ES2629353T3 - Individual variable domains of antibodies against serum albumin - Google Patents

Abstract

A ligand comprising an individual variable domain of antibodies, wherein said individual variable domain of antibodies comprises the amino acid sequence of dAb7h11 having the amino acid sequence depicted in Figure 24, Sheet 1, or an amino acid sequence that is identical to the same at least in 95%, where the beta T half-life of the individual variable domain in cynomolgus is at least 50% of the beta T half-life of cynomolgus serum albumin in cynomolgus.

Description

Individual variable domains of antibodies against serum albumin

The present invention relates to ligands comprising an individual variable domain of antibodies.

Introduction

The antigen binding domain of an antibody comprises two separate regions: a heavy chain variable domain (VH) and a light chain variable domain (VL: which can be Vκ or Vλ). The antigen binding site itself consists of six polypeptide bonds: three from the VH domain (H1, H2 and H3) and three from the VL domain (L1, L2 and L3). A diverse primary repertoire of V genes that encode the VH and VL domains is produced by combinatorial rearrangement of gene segments. The VH gene is produced by the recombination of three gene segments, VH, D

10 and JH. In humans, there are approximately 51 functional VH segments (Cook and Tomlinson (1995) Immunol. Today, 16: 237), 25 functional D segments (Corbett et al. (1997) J. Mol. Biol., 268: 69) and 6 functional JH segments (Ravetch et al. (1981) Cell, 27: 583), depending on the haplotype. The VH segment encodes the region of the polypeptide chain that forms the first and second antigen binding bonds of the VH domain (H1 and H2), while the VH, D and JH segments combine to form the third binding loop to VH domain antigen

15 (H3). The VL gene is produced by the recombination of only two gene segments, VL and JL. In humans, there are approximately 40 functional Vκ segments (Schäble and Zachau (1993) Biol. Chem. Hoppe-Seyler, 374: 1001), 31 functional Vλ segments (Williams et al. (1996) J. Mol. Biol., 264: 220; Kawasaki et al. (1997) Genome Res., 7: 250), 5 functional Jκ segments (Hieter et al. (1982) J. Biol. Chem., 257: 1516) and 4 functional Jλ segments (Vasicek and Leder (1990) J. Exp. Med., 172: 609), depending on the haplotype. The VL segment encodes the

20 region of the polypeptide chain that forms the first and second antigen binding bonds of the VL domain (L1 and L2), while the VL and JL segments combine to form the third antigen binding loop of the VL domain (L3) . The antibodies selected from this primary repertoire are thought to be sufficiently diverse to bind to almost all antigens with at least moderate affinity. High affinity antibodies are produced by "affinity maturation" of the rearranged genes, in which mutations are generated punctually and

25 are selected by the immune system based on their enhanced binding.

Analysis of antibody structures and sequences has shown that five of the six antigen-binding loops (H1, H2, L1, L2, L3) possess a limited number of main chain conformations or canonical structures (Chothia and Lesk (1987 ) J. Mol. Biol., 196: 901; Chothia et al. (1989) Nature, 342: 877). The main chain conformations are determined by (i) the length of the antigen binding loop, and (ii) 30 particular residues, or types of residue, at certain key positions in the antigen binding loop and the main structure of the antibody . The analysis of the loop lengths and key residues has allowed us to predict the main chain conformations of H1, H2, L1, L2 and L3 encoded by most human antibody sequences (Chothia et al. (1992) J. Mol. Biol., 227: 799; Tomlinson et al. (1995) EMBO J., 14: 4628; Williams et al. (1996) J. Mol. Biol., 264: 220). Although the H3 region is much more diverse in terms of sequence, length and

The structure (due to the use of segments D), also forms a limited number of main chain conformations for short loop lengths that depend on the length and presence of particular residues, or types of residue, at key positions in the loop and the main structure of the antibody (Martin et al. (1996) J. Mol. Biol., 263: 800; Shirai et al. (1996) FEBS Letters, 399: 1.

Bispecific antibodies comprising complementary pairs of VH and regions are known in the art.

40 VL These bispecific antibodies must comprise two pairs of VH and VL, with each VH / VL pair binding to a single antigen or epitope. The methods described involve hybridomas hybridomas (Milstein & Neck AC, Nature 305: 537-40), minibodies (Hu et al., (1996) Cancer Res 56: 3055-3061), diabodies (Holliger et al., (1993) Proc Natl. Acad. Sci. USA 90, 6444-6448; WO-94/13804), recombinant chelation antibodies (CRAb; (Neri et al., (1995) J. Mol. Biol. 246, 367-373) , biscFv (eg Atwell et al., (1996) Mol. Immunol. 33, 1301-1312), antibodies

45 stabilized "buttonhole" (Carter et al., (1997) Protein Sci. 6, 781-788). In each case each antibody species comprises two antigen binding sites, each adapted by a complementary pair of VH and VL domains. Each antibody is thus able to bind to two different antigens or epitopes at the same time, with binding to EACH antigen or epitope mediated by a VH domain and its complementary VL. Each of these techniques has its own drawbacks; for example in the case of hybrid hybridomas, the VH / VL pairs

50 inactive can greatly reduce the bispecific IgG fraction. In addition, most bispecific strategies depend on the association of different VH / VL pairs or the association of VH and VL chains to recreate the two different VH / VL binding sites. It is therefore impossible to control the ratio between binding sites to each antigen or epitope in the assembled molecule and thus many of the assembled molecules will bind to one antigen or epitope but not the other. In some cases it has been possible to design heavy or light chains in the

55 subunit interfaces (Carter et al., 1997) in order to improve the number of molecules that have binding sites for the two antigens or epitopes but this never happens in all molecules that have binding for the two antigens or epitopes .

There is some evidence that two different antibody binding specificities could be incorporated into the same binding site, but these generally represent two or more specificities that correspond to antigens or 60 epitopes or structurally related antibodies that are of very important cross-reactivity. By

For example, cross-reactivity antibodies have been described, usually in which the two antigens are related in sequence and structure, for example in the lysozyme of the chicken egg white or in the lysozyme of the turkey (McCafferty et al., WO document -92/01047) or with hapten and hapten conjugated to the vehicle (Griffiths AD et al. EMBO J 1994 13:14 3245-60). In a further example, WO-02/02773 (Abbott Laboratories) 5 describes "double specificity" antibody molecules. The aforementioned antibody molecules are antibodies prepared or selected against multiple antigens, so that their specificity extends to more than a single antigen. Each complementary VH / VL pair in the antibodies of WO-02/02773 specifies a single binding specificity for two or more structurally related antigens; the VH and VL domains in said complementary pairs do not each have a separate specificity. Thus the antibodies have a broad individual specificity comprising two antigens, which are structurally related. In addition, natural autoantibodies have been described that are polyreactive (Casali & Notkins, Ann. Rev. Immunol. 7, 515-531), which react with at least two (usually more) different antigens or epitopes that are not structurally related. It has also been shown that random peptide repertoire selections using phage display technology in a monoclonal antibody will identify a range of sequences.

15 peptides that conform to the antigen binding site. Some of the sequences are highly related, to fit a consensus sequence, while others are very different and have been called mimotopes (Track & Stephen, Current Opinion in Immunology, 1993, 5, 268-271). Therefore, it is clear that a natural four-chain antibody, which comprises associated and complementary VH and VL domains, has the potential to bind to many different antigens in a large universe of known antigens. It is not so clear how to create a binding site with two antigens given on the same antibody, particularly those that are not necessarily structurally related.

Protein design methods have suggested that they can offer support for this. For example it has also been proposed that a catalytic antibody could be created with a binding activity with a metal ion through a variable domain, and with a hapten (substrate) through contacts with the metal ion and a domain

25 complementary variable (Barbas et al., 1993 Proc. Natl. Acad. Sci USA 90, 6385-6389). However, in this case it is proposed that the binding and catalysis of the substrate (first antigen) requires the binding of the metal ion (second antigen). Thus the union with the VH / VL pair refers to a single but multi-component antigen.

Methods for the creation of bispecific antibodies from individual heavy chain domains of camel antibodies in which binding contacts are created for an antigen in a variable domain, and for a second antigen in a second variable domain have been described. However, the variable domains were not complementary. Thus a first heavy chain variable domain is selected against a first antigen, and a second heavy chain variable domain against a second antigen, and then the two domains are joined together in the same chain to provide a bispecific antibody fragment (Conrath et al., J. Biol. Chem. 270, 27589-27594). However, individual camel heavy chain domains are little

35 frequent since they come from natural camel antibodies that do not have light chains, and in fact individual heavy chain domains are unable to associate with camel light chains to form complementary VH and VL pairs.

Individual heavy chain variable domains, obtained from natural antibodies that are normally associated with light chains (from monoclonal antibodies or domain repertoires, have also been described; see EP-A-0368684). It has been shown that these heavy chain variable domains specifically interact with one or more related antigens, but have not been combined with other heavy or light chain variable domains to create a ligand with a specificity for two or more different antigens. In addition, it has been shown that these individual domains have a very short half-life in vivo. Therefore said domains have a limited therapeutic value.

It has been suggested to prepare bispecific antibody fragments by joining heavy chain variable domains of different specificity (as described above). The disadvantage of this strategy is that the variable domains of isolated antibodies can have a hydrophobic interface that normally interacts with the light chain and is indicated to the solvent and can be "sticky" which allows the individual domain to bind to hydrophobic surfaces. In addition, in the absence of an accompanying light chain the combination of two or more different heavy chain variable domains and their association, possibly through their hydrophobic interfaces, may prevent them from binding to one, if not the two ligands to which They could join in isolation. On the other hand, in this case the heavy chain variable domains would not be associated with complementary light chain variable domains and thus may be less stable and easy to deploy (Worn & Pluckthun, 1998 Biochemistry 37, 13120-7).

55 Summary of the invention

The inventors have described, in their international patent application in process with the present WO03 / 002609 as well as in the UK patent application not published in the process with the present 0230203.2, the double specific immunoglobulin ligands comprising domains individual immunoglobulin variables that each have different specificities. The domains can act in competition with each other or independently to bind antigens or epitopes on target molecules. WO2005 / 118642 describes individual variable domains that have binding specificity for serum albumin.

In a first aspect, the present invention provides a ligand comprising an individual variable domain of antibodies, wherein said individual variable domain of antibodies comprises the amino acid sequence of dAb7h14 having the amino acid sequence depicted in Figure 24, Sheet 1, or an amino acid sequence that is identical to it at least 95%, where the beta T half-life of the variable domain

5 individual in cynomolgus is at least 50% of the beta T half-life of cynomolgus serum albumin in cynomolgus.

The present invention also provides a ligand as described above, wherein the individual variable domain of antibodies is bound to rat, mouse, human serum albumin and Cynomolgus.

The present invention also provides a ligand as described above, wherein said domain

The individual variable antibody is specifically bound to human serum albumin and non-human serum albumin with Kd values within the limits of a factor of 10 to each other.

The present invention also provides a ligand as described above, further comprising a biologically active polypeptide.

The present invention also provides a ligand as described above, wherein said individual variable domain of antibodies is conjugated to a drug.

An additional improvement in specific double ligands as described by the authors of the present invention is described herein, in which a specificity of the ligand is directed towards a protein or polypeptide present in vivo in an organism that can act to increase the half-life of the ligand joining him.

Accordingly, a double specific ligand comprising a first individual variable domain of immunoglobulins having a specificity of binding to a first antigen or epitope and a second individual variable domain of complementary immunoglobulins having a binding activity is described herein. to a second antigen or epitope, where one or both of said antigens or epitopes act to increase the half-life of the ligand in vivo and where said first and second domains lack mutually domains.

Complementary compounds that share the same specificity, provided that said double specific ligand does not consist of an anti-HSA VH domain and an anti-galactosidase Vκ domain. Preferably, none of the first or second variable domains is bound to human serum albumin (HSA).

Antigens or epitopes that increase the half-life of a ligand as described herein are advantageously present in proteins or polypeptides present in an organism in vivo. Examples include

30 extracellular matrix proteins, blood proteins and proteins present in various tissues in the body. The proteins act to reduce the clearance rate of the blood ligand, for example by acting as bulking agents, or by anchoring the ligand to a desired site of action. Examples of antigens / epitopes that increase half-life in vivo are provided in Annex 1 below.

The half-life increase is useful in in vivo applications of immunoglobulins, especially antibodies and very

35 especially small antibody fragments. These fragments (Fv, Fv with disulfide bonds, Fab, scFv, dAb) undergo rapid clearance of the organism; Thus, although they are able to reach most parts of the body quickly, and are quick to produce and easy to handle, their in vivo applications have been limited by their brief persistence in vivo. The invention solves this problem by providing a longer half-life of the ligands in vivo and therefore longer persistence times in the body of

40 functional activity of the ligand.

The methods for pharmacokinetic analysis and determination of ligand half-life will be familiar to those skilled in the art. Details of them can be found in Kenneth, A et al: Chemical Stability of Pharmaceuticals: A Handbook to Pharmacists and in Peters et al., Pharmacokinetc analysis: A Practical Approach (1996). Reference is also made to "Pharmacokinetics", M Gibaldi & D Perron, published by Marcel Dekker, 2nd

45 edition rev. ex (1982), which describes pharmacokinetic parameters such as t alpha and t beta half-lives and the area under the curve (ABC).

Half-lives (t½ alpha and t½ beta) and ABC can be determined from a serum ligand concentration curve with respect to time. To model the curve, for example, the WinNonlin analysis package (available from Pharsight Corp., Mountain View, CA94040, USA) can be used. In a first phase (the alpha phase) the ligand is mainly experiencing distribution in the patient, with some elimination. A second phase (beta phase) is the terminal phase when the ligand has been distributed and the serum concentration is decreasing as the ligand clears in the patient. The half-life alpha t is the half-life of the first phase and the half-life beta t is the half-life of the second phase. A ligand or a composition comprising a ligand having a half-life tα in the range of 15 minutes or more is described herein. In one embodiment, the lower end of the interval is 30 minutes, 45 minutes, 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 10 hours, 11 hours

or 12 hours In addition, or alternatively, a ligand or composition having a half-life tα in the range, up to 12 hours, is described herein. In one embodiment, the upper end of the range is 11, 10, 9, 8, 7, 6 or 5 hours. An example of a suitable interval is 1 to 6 hours, 2 to 5 hours or 3 to 4 hours.

A ligand or a composition comprising a ligand having a half-life tβ in the range of 2.5 hours or more is described herein. In one embodiment, the lower end of the interval is 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 10 hours, 11 hours or 12 hours. In addition, or alternatively, a ligand or composition has a tβ half-life in the range, inclusive, of up to 21 days. In one embodiment, the end

Top 5 of the interval is 12 hours, 24 hours, 2 days, 3 days, 5 days, 10 days, 15 days or 20 days. Advantageously a ligand or composition will have a tβ half-life in the range of 12 to 60 hours. In a further embodiment, it will be in the range of 12 to 48 hours. In a further embodiment, it will be in the range of 12 to 26 hours.

In addition, or alternatively to the above criteria, the description provides a ligand or a composition comprising a ligand having an ABC value (area under the curve) in the range of 1 mg / min / ml or more. In a

10 embodiment, the lower end of the range is 5, 10, 15, 20, 30, 100, 200 or 300 mg / min / ml. In addition, or alternatively, a ligand or composition has an ABC in the range of up to 600 mg / min / ml. In one embodiment, the upper end of the range is 500, 400, 300, 200, 150, 100, 75 or 50 mg / min / ml. Advantageously, a ligand will have an ABC in the range selected from, but preferably not limited to, the group consisting of the following: 15 to 150 mg / min / ml, 15 to 100 mg / min / ml, 15 to 75 mg / min / ml and 15 to 50 mg / min / ml.

In a first embodiment, the double specific ligand comprises two complementary variable domains, that is, two variable domains that, in their natural environment, are capable of operating together as a similar pair or group, even if in the context of the present invention. they bind separately to their similar epitopes. For example, complementary variable domains may be heavy chain and light chain (VH and VL) variable domains of immunoglobulins. The VH and VL domains are advantageously provided by fragments of

20 scFv or Fab antibodies. Variable domains can be linked together to form multivalent ligands, for example: by providing a hinge region at the C end of each V domain and disulfide bonds between cysteines in the hinge regions; or the provision of dAb each with a cysteine at the C-terminus of the domain, the cysteines being linked to each other by disulfides; or the production of V-CH and V-CL to produce a Fab format; or the use of peptide linker groups (eg, Gly4Ser linker groups discussed later in the

25 present memory) to produce additional dimers, trimers and multimers.

The authors of the invention have discovered that the use of complementary variable domains allows the joint packaging of the two domain surfaces and their solvent sequestration. In addition, complementary domains are able to stabilize each other. In addition, this allows the creation of double specific IgG antibodies without the disadvantages of hybrid hybridomas as used in the prior art, or without the need to design heavy or light chains at the subunit interfaces. The specific double ligands described herein have at least one VH / VL pair. A bispecific IgG described herein will therefore comprise two of these pairs, a pair in each arm of the Y-shaped molecule. Therefore, unlike the bispecific antibodies or conventional diabodies, in which the ratio between the chains used is decisive in the success of their preparation and leads to practical difficulties, the double specific ligands described herein are exempt from chain balance problems. The chain imbalance in conventional bispecific antibodies comes from the association of two different VL chains with two different VH chains, in which the VL 1 chain together with the VH 1 chain is capable of binding to the antigen or epitope 1 and the VL chain 2 together with the VH 2 chain is capable of binding to the antigen or epitope 2 and the two correct pairs are somehow linked to each other. Thus, only when the VL 1 chain is paired with the VH 1 chain and the VL 2 chain is paired with the VH 2 chain in a single molecule is the bispecificity created. Such bispecific molecules can be created in two different ways. First, they can be created by association of two existing VH / VL pairs that each bind to a different antigen or epitope (for example, in a bispecific IgG). In this case the VH / VL pairs must be together in a 1: 1 ratio in order to create a population of molecules all of which are bispecific. This never happens (even when the complementary CH domain is reinforced by a "buttonhole" type design) which leads to a mixture of bispecific molecules and molecules that can only bind to one antigen or epitope but not the other. The second way to create a bispecific antibody is by simultaneously associating two different VH chains with two different VL chains (for example in a bispecific diabody). In this case, although there is usually a preference for VL 1 chains to pair with VH 1 chain and VL 2 chain to pair with

50 the VH 2 chain (which can be reinforced by a "buttonhole" type design of the VL and VH domains), this pairing is never achieved in all molecules, which leads to a mixed formulation and thus to production of mismatches that are unable to bind to the antigen or epitope.

Bispecific antibodies constructed according to the double specific ligand strategy described herein overcome all these problems because the binding to the antigen or epitope 1 resides in the VH or VL domain and the binding

The antigen or epitope 2 resides in the complementary VL or VH domain, respectively. Since the VH and VL domains are paired on a 1: 1 basis all VH / VL pairings will be bispecific and thus all formats constructed using these VH / VL pairings (Fv, scFv, Fab, minibodies, IgG, etc.). They will have a 100% bispecific activity.

In the context of the present description, it is understood that the first and second "epitopes" are epitopes that are not

60 are the same and are not bound by an individual monospecific ligand. In the first configuration, they are advantageously in different antigens, one of which acts to increase the half-life of the ligand in vivo. Similarly, the first and second antigens are advantageously not the same.

The specific double ligands described herein do not include ligands as described in WO-02/02773. Thus, the ligands described herein do not comprise complementary VH / VL pairs that bind to any one or more antigens or epitopes cooperatively. Instead, the ligands described herein comprise a complementary VH / VL pair, wherein the V domains

5 have different specificities.

On the other hand, the ligands described herein comprise complementary VH / VL pairs that have different specificities for epitopes or non-structurally related antigens. Structurally related epitopes or antigens are epitopes or antigens that possess sufficient structural similarity to be linked by a conventional complementary VH / VL pair that acts cooperatively to bind to an antigen or

10 epitope; in the case of structurally related epitopes, the epitopes are sufficiently similar in structure to "fit" into the same binding bag formed at the antigen binding site of the VHNL dimer.

In one aspect, the description provides a ligand comprising a first immunoglobulin variable domain that has a first antigen or epitope binding specificity and a second immunoglobulin variable domain that has a second antigen or epitope binding specificity, wherein one or both of

15 said first and second variable domains bind to an antigen that increases the half-life of the ligand in vivo, and the variable domains are not complementary to each other.

In one embodiment, binding to a variable domain modulates the binding of the ligand to the second variable domain.

In this embodiment, the variable domains can be, for example, pairs of VH domains or pairs of VL domains. Antigen binding at the first site can modulate, for example improve or inhibit, the binding of an antigen at the site.

20 second site. For example, binding at the first site at least partially inhibits the binding of an antigen at a second site. In said embodiment, the ligand can be maintained, for example, in the body of an object organism in vivo through binding to a protein that increases the half-life of the ligand until such time as it binds to the second target antigen and dissociates from the protein. It increases protein.

The modulation of the union in the previous context is achieved as a consequence of the structural proximity

25 of the relative antigen binding sites between them. Such structural proximity can be achieved from the nature of the structural components that bind the two or more antigen binding sites, e.g. eg, by providing a ligand with a relatively rigid structure that keeps antigen binding sites in close proximity. Advantageously, the two or more antigen binding sites are in close physical proximity to each other so that one site modulates antigen binding at another site by a process that

30 implies steric hindrance and / or conformational changes in the immunoglobulin molecule.

The first and second antigen binding domains can be covalently or non-covalently associated. In the case that the domains are covalently associated, the association can be mediated for example by disulfide bonds or by a polypeptide linker group such as (Gly4Ser) n, in which n = from 1 to 8, p. e.g., 2, 3, 4, 5 or

7.

The ligands according to the invention can be combined into multiple non-immunoglobulin ligand structures to form multivalent complexes, which bind to target molecules with the same antigen, thus providing greater avidity, while at least one variable domain binds to an antigen. to increase the half-life of the multimer. For example, natural bacterial receptors such as SpA have been used as frameworks for CDR grafting to generate ligands that specifically bind to one or more epitopes. The details of this

The procedure is described in US 5,831,012. Other suitable frameworks include those based on fibronectin and affibody ™ molecules. Details of suitable procedures are described in WO98 / 58965. Other suitable frameworks include lipocalin and CTLA4, as described in van den Beuken et al., J. Mol. Biol. (2001) 310, 591-601, and frameworks such as those described in WO0069907 (Medical Research Council), which are based for example on the ring structure of bacterial GroEL or other polypeptides of

45 chaperones

Protein frameworks can be combined; for example, CDRs can be grafted onto a CTLA4 framework and used together with VH or VL immunoglobulin domains to form a ligand. Similarly, fibronectin, lipocalin and other frameworks can be combined.

In the case that the variable domains are selected among repertoires of selected V genes for example

50 using phage display technology as described herein, then these variable domains may comprise a universal framework region, so that they can be recognized by a specific generic ligand as defined herein. The use of universal frames, generic ligands and the like is described in WO99 / 20749. In the present description, the reference to phage display includes the use of phages and / or phagemids.

When V gene repertoires are used, the variation in the polypeptide sequence is preferably located in the structural bonds of the variable domains. The polypeptide sequences of any variable domain can be modified by shuffling DNA or by mutation in order to enhance the interaction of each variable domain with its complementary pair.

The ’double specific ligand’ can be a single chain Fv fragmenter that can consist of an Fab region of an antibody. The expression "Fab region" includes a region of type Fab in which two VH domains or two VL domains are used.

Variable domains can be obtained from antibodies directed against target antigens or epitopes. 5 Alternatively, they can be obtained from a repertoire of individual antibody domains such as those expressed on the surface of filamentous bacteriophages. The selection can be made as described below.

A method for producing a ligand comprising a first individual variable domain of immunoglobulins having a first binding specificity and a second individual variable domain of immunoglobulins having a second binding specificity (different) is described herein, being one or both

10 specific binding specificities of an antigen that increases the half-life of the ligand in vivo, the method comprising the steps of:

(to)

 selection of a first variable domain for its ability to bind to a first epitope,

(b)

 selection of a second variable domain for its ability to join a second epitope,

(C)

 combination of variable domains; Y

15 (d) ligand selection for its ability to bind said first epitope and said second epitope.

The ligand can bind to the first and second epitopes either simultaneously or, when there is competition between binding domains for epitope binding, the binding of one domain can prevent the binding of another domain to its similar epitope. In one embodiment, therefore, step (d) above requires simultaneous binding to the first and second (and possibly more) epitopes; in another embodiment, binding to the first and second epitopes does not

20 is simultaneous.

The epitopes are preferably in separate antigens.

The ligands advantageously comprise VH / VL combinations, or VH / VH or VL / VL combinations of immunoglobulin variable domains, as described above. On the other hand, the ligands may comprise VHH domains of camelids, provided that the VHH domain that is specific for an antigen that increases the

The ligand half-life in vivo does not bind to a lysozyme from chicken egg white (HEL), a porcine pancreatic alpha-amylase or NmC-A; hcg, RR6 azo dye bound to BSA or S. mutans HG982 cells, as described in Conrath et al., (2001) JBC 276: 7346-7350 and in WO99 / 23221, none of which describes the use of a specificity for an antigen that increases the half-life to increase the half-life of the ligand in vivo.

In one embodiment, said first variable domain is selected to bind to said first epitope in the absence of

30 a complementary variable domain. In a further embodiment, said first variable domain is selected to bind to said first epitope / antigen in the presence of a third variable domain in which said third variable domain is different from said second variable domain and is complementary to the first domain. Similarly, the second domain can be selected in the absence or presence of a complementary variable domain.

Antigens or epitopes directed by the ligands described herein that increase the

Half-life of a ligand is not limited to serum albumin targets. Other embodiments of antigens or epitopes directed by the ligands described herein that increase the half-life of a ligand in vivo include, but are preferably not limited to, the antigens and epitopes listed in Annex 1 below.

The antigens or epitopes directed by the ligands of the invention, in addition to the protein that improves the

Half-life, can be any antigen or epitope, although advantageously they are antigen or epitope directed with a therapeutic benefit. The invention provides ligands, which include open conformation, closed conformation and ligands of isolated dAb monomers, specific for any of said targets, in particular the targets identified especially herein. Said targets may be, or be part of, polypeptides, proteins or nucleic acids, which may be of natural or synthetic occurrence. In this regard, the

The ligand of the invention can bind to the epitope or antigen and act as an antagonist or agonist (eg, EPO receptor agonist). One skilled in the art will observe that the choice is wide and varied. These can be, for example, proteins, cytokines, human or animal cytokine receptors, in which cytokine receptors include receptors for cytokines, enzymes, cofactors for enzymes or DNA binding proteins. Suitable cytokines and growth factors include, but preferably are not limited to: ApoE, Apo-SAA, BDNF,

50 Cardiotrophin-1, EGF, EGF receptor, ENA-78, Eotaxin, Eotaxin-2, Exodus-2, EpoR, FGF acid, basic FGF, fibroblast growth factor-10, ligand FLT3, Fractalcin (CX3C), GDNF , G-CSF, GM-CSF, GF-β1, insulin, IFN-γ, IGF-I, IGF-II, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5 , IL-6, IL-7, IL-8 (72 aa)., IL-8 (77 aa)., IL-9, IL-10, IL11, IL-12, IL-13, IL-15, IL -16, IL-17, IL-18 (IGIF), Inhibin α, Inhibin β, IP-10, keratinocyte growth factor2 (KGF-2), KGF, Leptin, LIF, Lympotactin, Müller inhibitor substance, inhibitory factor of monocyte colonies,

Monocyte attraction protein, M-CSF, MDC (67 aa.), MDC (69 aa)., MCP-1 (MCAF), MCP-2, MCP-3, MCP-4, MDC (67 aa). , MDC (69 aa)., MIG, MIP-1α, MIP-1β, MIP-3α, MIP-3β, MIP-4, parent inhibitor factor

Myeloids-1 (MPIF-1), NAP-2, Neurturin, Nerve Growth Factor, β-NGF, NT-3, NT-4, Oncostatin M, PDGF-AA, PDGF-AB, PDGF-BB, PF-4 , RANTES, SDF1α, SDF1β, SCF, SCGF, stem cell factor (SCF), TARC, TGF-α, TGF-β, TGF-β2, TGF-β3, tumor necrosis factor (TNF), TNF-α, TNF -β, TNF receptor I, TNF II receptor, TNIL-1, TPO, VEGF, VEGF 1 receiver, VEGF 2 receiver, VEGF 3 receiver, GCP-2, GRO / MGSA, GRO-β,

5 GRO-γ, HCC1, 1-309, HER 1, HER 2, HER 3 and HER 4, CD4, human chemokine receptors CXCR4 or CCR5, non-structural protein type 3 (NS3) of hepatitis C virus, TNF- alpha, IgE, IFN-gamma, MMP-12, CEA, H. pylori, TB, influenza, Hepatitis E, MMP-12, internalization receptors that are overexpressed in certain cells, such as the epidermal growth factor receptor (EGFR) , ErBb2 receptor in tumor cells, an internalization cell receptor receptor, LDL receptor, FGF2 receptor, ErbB2 receptor, transferrin receptor, PDGF receptor, VEGF receptor, PsmAr, an extracellular matrix protein, elastin, fibronectin, laminin, α1-antitrypsin , tissue factor protease inhibitor, PDK1, GSK1, Bad, caspase-9, Forkhead, a Helicobacter pylori antigen, a Mycobacterium tuberculosis antigen and a influenza virus antigen. It will be noted that this list is by no means exhaustive.

In one embodiment of the invention, the variable domains are derived from a respective antibody directed against the

15 antigen or epitope. In a preferred embodiment the variable domains come from a repertoire of individual variable antibody domains.

In a further aspect, the present disclosure provides one or more nucleic acid molecules that encode at least one specific double ligand as defined herein. The specific double ligand may be encoded in a single nucleic acid molecule; alternatively, each domain may be encoded by an independent nucleic acid molecule. When the ligand is encoded by a single nucleic acid molecule, the domains can be expressed as a fusion polypeptide, in the manner of a scFv molecule, or they can be expressed separately and subsequently linked, for example using chemical binding agents. Ligands expressed from separated nucleic acids will be linked together by suitable means.

The nucleic acid can also encode a signal sequence for the export of the polypeptides from

A host cell after expression and can be fused with a surface component of a filamentous bacteriophage particle (or another component of a selection presentation system) after expression.

The specific double ligands described herein preferably comprise combinations of heavy and light chain domains. For example, the specific double ligand may comprise a VH domain and a VL domain, which may be linked together in the form of an scFv. In addition, the ligands may comprise one

or more domains CH or CL. For example, the ligands may comprise a CH1 domain, a CH2 or CH3 domain and / or a CL domain, C1, C2, C3 or C4 domains, or any combination thereof. A hinge region domain can also be included. Such combinations of domains can mimic, for example, natural antibodies, such as IgG or IgM, or fragments thereof, such as Fv, scFv, Fab or F (ab ’) 2 molecules. Other structures are also contemplated, such as an individual arm of an IgG molecule comprising VH, VL, CH1 and CL domains.

35 Variable regions can be selected from repertoires of V genes from individual domains. In general the repertoire of individual antibody domains occurs on the surface of filamentous bacteriophages. In a preferred embodiment each individual antibody domain is selected by binding a phage repertoire to the antigen.

Each individual variable domain can be selected for binding to its target antigen or epitope in the absence of a complementary variable region. In an alternative embodiment, the individual variable domains may be selected for binding to their target antigen or epitope in the presence of a complementary variable region. Thus, the first individual variable domain can be selected in the presence of a third complementary variable domain, and the second variable domain can be selected in the presence of a fourth complementary variable domain. The third or fourth complementary variable domain may be in the like variable domain

Natural that has the same specificity as the individual domain under test, or a complementary non-similar domain, as a "training" variable domain.

Preferably, the double specific ligand described herein comprises only two variable domains although it is possible to incorporate several of these ligands together in the same protein, for example two of these ligands can be incorporated into an IgG or a multimeric immunoglobulin, such as IgM. Alternatively, in another embodiment a plurality of specific double ligands is combined to form a multimer. For example, two different double specific ligands are combined to create a tetra-specific molecule.

One skilled in the art will note that the light and heavy variable domains of a specific double ligand produced according to the method described herein may be in the same polypeptide chain, or alternatively, in different polypeptide chains. In the case where the variable domains are in strings

From different polypeptides, they can be linked by means of a linker group, in general a flexible linker group (such as a polypeptide chain), a chemical linker group or any other method known in the art.

A composition comprising a specific double ligand is described herein, which may

Obtained by a method described herein, and a pharmaceutically acceptable carrier, diluent or excipient.

On the other hand, a method for the treatment and / or prevention of diseases using a "specific double ligand" or a composition described herein is described herein.

5 In a second configuration, the description provides multispecific ligands comprising at least two non-complementary variable domains. For example, the ligands may comprise a pair of VH domains or a pair of VL domains. Advantageously, the domains have a non-camelid origin; preferably they are human domains or comprise human framework regions (FW) and one or more heterologous CDRs. CDRs and framework regions are those regions of a variable domain of immunoglobulins as defined in the Kabat database.

10 Sequences of Proteins of Immunological Interest.

Preferred human framework regions are those encoded by germline gene segments DP47 and DPK9. Advantageously, FW1, FW2 and FW3 of a VH or VL domain have the sequence of FW1, FW2 or FW3 of DP47 or DPK9. Human frames may optionally contain mutations, for example up to about 5 amino acid changes or up to about 10 amino acid changes collectively in the frames.

15 humans used in the ligands of the invention.

Variable domains in multispecific ligands according to the description herein may be organized in an open or closed conformation; that is, they can be arranged so that the variable domains can bind to their similar ligands independently and simultaneously, or so that only one of the variable domains can bind to their similar ligand at any given time.

The authors of the invention have understood that under certain structural conditions, non-complementary variable domains (for example two light chain variable domains or two heavy chain variable domains) may be present in a ligand so that the binding of a first epitope to a first variable domain inhibits the binding of a second epitope to a second variable domain, even if said non-complementary domains do not act together as a similar pair. Advantageously, the ligand comprises two or

25 more pairs of variable domains; that is, it comprises at least four variable domains. Advantageously, the four variable domains comprise frames of human origin.

In a preferred embodiment, the human frames are identical to those of the human germline sequences.

The authors of the present invention consider that said antibodies will have a particular use in ligand binding assays for therapeutic and other uses.

Thus, the description provides a method for producing a multispecific ligand comprising the steps of:

a) selection of a first epitope binding domain for its ability to bind to a first epitope,

b) selection of a second epitope binding domain for its ability to bind to a second epitope,

c) combination of epitope binding domains; Y

d) selection of the multispecific conformal ligand closed for its ability to bind to said first second epitope and said second epitope.

In a further aspect of the second configuration, the description provides a method for preparing a multi-specific closed-forming ligand comprising a first epitope binding domain that has a first epitope binding specificity and a second complementary epitope binding domain that it has a second epitope binding specificity, where the first and second binding specificities compete for

The epitope binding so that the multi-specific closed conformation ligand cannot bind to the two epitopes simultaneously, said method comprising the steps of:

a) selection of a first epitope binding domain for its ability to bind to a first epitope,

b) selection of a second epitope binding domain for its ability to bind to a second epitope,

c) combination of epitope binding domains so that the domains are in a closed conformation; Y

d) selection of the multispecific conformation ligand closed for its ability to bind to said first second epitope and said second epitope, but not to said first and second epitopes simultaneously.

On the other hand, the description provides a multispecific closed conformation ligand comprising a first epitope binding domain that has a first epitope binding specificity and a second non-complementary epitope binding domain 50 that has a second binding specificity. epitopes, where the first and second binding specificities compete for epitope binding so that the ligand

Multi-specific closed conformation cannot bind to both epitopes simultaneously.

An alternative embodiment of the above aspect of the second configuration of the description optionally comprises an additional step (b1) comprising the selection of a third or more epitope binding domain. In this way the multispecific ligand produced, whether it has open or closed conformation, comprises

5 more than two epitope binding specificities. In a preferred aspect of the second configuration of the description, in which the multispecific ligand comprises more than two epitope binding domains, at least two of said domains are in a closed conformation and compete for binding; other domains may compete for union or may be free to associate independently with their epitope or similar epitopes.

The term "multispecific ligand" refers to a ligand that has more than one specificity of binding to 10 epitopes as defined herein.

As defined herein the term "closed conformation" (multispecific ligand) means that the epitope binding domains of the ligand are linked or associated with each other, optionally by means of a protein skeleton, so that epitope binding for an epitope binding domain competes with epitope binding for another epitope binding domain. That is, similar epitopes can be linked by each

15 epitope binding domain individually, but not simultaneously. The closed conformation of the ligand can be achieved using the methods described herein.

"Open conformation" means that the epitope binding domains of the ligand are linked or associated with each other, optionally by means of a protein skeleton, such that epitope binding by an epitope binding domain does not compete with binding to epitopes by another epitope binding domain.

20 As referred to herein, the term "competes" means that the binding of a first epitope to its similar epitope binding domain is inhibited when a second epitope binds to its similar epitope binding domain. For example, binding can be sterically inhibited, for example by physical blocking of a binding domain or by altering the structure or environment of a binding domain so that its affinity or greed for an epitope is reduced.

In a further embodiment of the second configuration, epitopes can move mutually in the junction. For example, a first epitope may be present in an antigen that, by binding to its first similar binding domain, causes a steric hindrance of a second binding domain, or a conformational change therein, which displaces the epitope attached to the second binding domain.

Advantageously, the union is reduced by 25% or more, advantageously 40%, 50%, 60%, 70%, 80%, 90

30% or more, and preferably up to 100% or about, so that the binding is completely inhibited. Epitope binding can be measured by conventional antigen binding assays, such as ELISA, by fluorescence-based techniques, including FRET, or by techniques such as resonance by surface plasmons that measure the mass of the molecules. Specific binding of an antigen binding protein with an antigen or epitope can be determined by a suitable assay, which includes, for example, Scatchard analysis and / or binding assays.

35 competitive, such as radioimmunoassays (RIA), enzymatic immunoassays such as ELISA and intercalation competition tests, and the different variants thereof.

Binding affinity is preferably determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden). The Biacore system uses surface plasmon resonance (SPR, Welford K. 1991, Opt. Quant. Elect. 23: 1; Morton and Myszka, 1998, Methods in 40 Enzymology 295: 268) to monitor biomolecular interactions in real time, and employs surface plasmon resonance that can detect changes in the resonance angle of light on the surface of a thin gold film on a glass support as a result of changes in the refractive index of the surface until a separation of 300 nm Biacore analysis conveniently generates association rate constants, dissociation rate constants, equilibrium dissociation constants and affinity constants. Binding affinity is obtained by assessing the constants of association and dissociation rates by using a Biacore surface plasmon resonance system (Biacore, Inc.). A biosensors chip for covalent coupling of the target is activated according to the manufacturer's instructions (Biacore). The target is then diluted and injected into the chip to obtain a signal in response units of immobilized material. Since the signal in resonance units (RU) is proportional to the mass of immobilized material,

This represents a range of immobilized target densities in the matrix. Dissociation data fit a single site model to obtain koff +/- d.t. (standard deviation of measures). A pseudo-first order velocity constant (Kd) is calculated for each association curve, and plotted as a function of protein concentration to obtain kon +/- e.s. (standard error of adjustment). The equilibrium dissociation constants for the junction, Kd, are calculated from SPR measurements such as koff / kon.

55 As described by Phizicky and Field in Microb. Rev. (1995) 59: 114-115, a suitable antigen, such as HSA, is immobilized in a dextran polymer, and a solution containing a ligand for HSA, as an individual variable domain, flows through a cell, in contact with the immobilized HSA. The individual variable domain retained by the immobilized SHA alters the resonance angle of the incident light, which causes a change in the rate of

refraction caused by the increase in protein amounts, that is, the individual variable domain, near the dextran polymer. Since all proteins have the same refractive index and since there is a linear correlation between the displacement of the resonance angle and the concentration of proteins near the surface, changes in the concentration of surface proteins due to binding can be measured.

5 protein / protein, see Phizicky and Field, above. To determine a binding constant, the increase in resonance units (RU) is measured as a function of time by passing a solution of individual variable domain proteins through the immobilized ligand (HSA) until the RU values are stabilized, and then The decrease in RU is measured as a function of time with buffer that lacks the individual variable domain. This procedure is repeated for several different concentrations of individual variable domain proteins. The theoretical background and detailed procedures are described in R. Karlsson, et. to the. (991) J. Immunol. Methods, 145, 229.

The instrument software produces an equilibrium dissociation constant (Kd) as described above. A equilibrium dissociation constant determined by the use of surface plasmon resonance is described in US Pat. 5,573,957, based on a table of values of dRA / dt and RA,

15 in which R in this example is the HSA / individual variable domain complex measured by Biacore in resonance units and in which dR / dt is the formation rate of HSA / individual variable domain complexes, that is, the derivative of the bonding curve; representing the graph dRA / dt with respect to RA for several different concentrations of individual variable domain, and subsequently representing the slopes of these lines with respect to the concentration of individual variable domain, the slope of this second graph is the association rate constant (M-1, s-1). The dissociation rate constant or the rate at which the HSA and the individual variable domain release each other can be determined using the dissociation curve generated in Biacore. Graphing and determining the slope of the log of the decrease in the response with respect to the time curve, the dissociation rate constant can be measured. Equilibrium dissociation constant Kd = Dissociation rate constant / Association velocity constant.

According to the method described herein, advantageously, each individual variable epitope binding domain has a different epitope binding specificity.

In the context of the present description, it is understood that the first and second "epitopes" are epitopes that are not the same and are not bound by a single monospecific ligand. They may be in different antigens or in the same antigen, but separated by a sufficient distance so that they do not form a single entity that could be linked by a monospecific VH / VL binding pair of a conventional antibody. Experimentally, if the two individual variable domains in a single form of chain antibody (domain antibodies or dAb) compete separately for a monospecific VH / VL ligand against two epitopes then these two epitopes are not sufficiently separated to be considered independent epitopes. .

The multi-specific closed conformation ligands described herein do not include ligands such as

35 is described in WO-02/02773. Thus, the ligands described herein do not comprise complementary VH / VL pairs that bind to any one or more antigens or epitopes cooperatively. Instead, the ligands described herein preferably comprise non-complementary VH-VH or VL-VL pairs. Advantageously, each VH or VL domain in each VH-VH or VL-VL pair has a different epitope binding specificity, and epitope binding sites are arranged such that the binding of an epitope at a site competes with the binding of an epitope somewhere else.

Advantageously, each epitope binding domain comprises a variable domain of immunoglobulins. More advantageously, each epitope binding domain will be a light chain variable domain (VL) or a heavy chain variable domain (VH) of an antibody. In the second configuration of the present description, the immunoglobulin domains when present in a ligand according to the present invention are not

Complementary, that is, do not associate to form a VH / VL antigen binding site. Thus, multispecific ligands as defined herein comprise immunoglobulin domains of the same subtype, that is, light chain variable domains (VL) or heavy chain variable domains (VH). On the other hand, when the ligand according to the invention is in closed conformation, the immunoglobulin domains can be of the VHH type of camelid.

In an alternative embodiment, the ligand (s) according to the invention do not comprise a camelid VHH domain. More specifically, the ligand (s) of the invention do not comprise one or more amino acid residues that are specific to camelid VHH domains compared to human VH domains.

Advantageously, the individual variable domains are obtained from antibodies selected for binding activity against different antigens or epitopes. For example, variable domains can be isolated at least in

55 part by human immunization. Alternative methods are known in the art, including isolation from human antibody libraries and synthesis of artificial antibody genes.

In selected embodiments an individual variable domain is an individual variable domain of natural occurrence. In other selected embodiments, the individual variable domain is of an unnatural occurrence. The term "natural occurrence" is used herein to indicate that an object, e.g. e.g., a protein domain,

p. eg, an individual variable domain, or an individual variable domain of antibodies, can be found in nature. Thus, a naturally occurring protein domain, such as a V region of an antibody, exists in a protein, e.g. ex. in an antibody chain protein, expressed in nature, for example, in a non-recombinant species, e.g. eg, mammals, primates, rodents, fish, birds, reptiles, etc. To rule out doubts, a domain

The individual variable isolated from a repertoire of polypeptides expressed from nucleic acids in which diversity was introduced in vitro is a single variable domain of unnatural occurrence. To rule out further doubts, an individual variable domain of antibodies that comes from an antibody resulting from immunization of an animal is an individual variable domain of natural occurrence.

Variable domains advantageously bind to superantigens, such as protein A or protein L. The binding to

10 superantigens is a property of the variable antibody domains with correct folding, and allows such domains to be isolated, for example, from recombinant or mutant domain libraries.

The epitope binding domains according to the present invention comprise a protein framework and epitope interaction sites (which are advantageously on the surface of the protein framework).

Epitope binding domains can also be based on scaffolds or skeletons of proteins other than

15 domains of immunoglobulins. For example, natural bacteria receptors such as SpA have been used as frameworks for CDR grafting in order to generate ligands that specifically bind to one or more epitopes. The details of this procedure are described in US-5,831,012. Other suitable frameworks include those based on fibronectin and affibody molecules. Details of suitable procedures are described in WO-98/58965. Other suitable frameworks include lipocalin and CTLA4, as

20 described in van den Beuken et al., J. Mol. Biol. (2001) 310, 591-601, and frames as described in WO0069907 (Medical Research Council), which are based, for example, on the ring structure of bacterial GroEL or other chaperone polypeptides.

Protein frameworks can be combined; for example, CDRs can be grafted onto a CTLA4 framework and used together with VH or VL immunoglobulin domains to form a multivalent ligand.

Similarly, fibronectin, lipocalin and other frameworks can be combined.

One skilled in the art will observe that the epitope binding domains of a multispecific closed conformation ligand produced according to the method described herein may be in the same polypeptide chain, or alternatively, in different polypeptide chains. In the case that the variable domains are in different polypeptide chains, they can be linked by means of a linker group, advantageously

A flexible linker group (such as a polypeptide chain), a chemical linker group or any other method known in the art.

The first and second epitope binding domains can be associated by covalent or non-covalent means. In the event that the domains are covalently associated, the association may be mediated for example by disulfide bonds.

In the second configuration, the first and second epitopes are preferably different. They can be, or be part of, polypeptides, proteins or nucleic acids, which can be of natural or synthetic occurrence. In this regard, the ligand can bind to an epitope or antigen and act as an antagonist or agonist (eg, EPO receptor agonist). The epitope binding domains of the ligand in one embodiment have the same epitope specificity, and for example they can simultaneously bind to their epitope when they are in the same antigen.

40 present multiple copies of the epitope. In another embodiment, these epitopes are provided in different antigens so that the ligand can bind to the epitopes and bypass the antigens. One skilled in the art will observe that the amount of epitopes and antigens is as large and varied as any target described in Annex 2 or Annex 3, either in combination as indicated in the Annexes, in a different combination or individually. Cytokine receptors include receptors for the above cytokines, e.g. ex. IL-1 R1; IL-6R; IL-10R; IL-18R as well

45 as receptors for the cytokines set out in Annex 2 or Annex 3 and also receivers described in Annexes 2 and 3. It will be noted that this list is by no means exhaustive. When the multispecific ligand is bound to two epitopes (in the same antigen or in different antigens), the antigen (s) can be selected from this list.

Advantageously, the specific double ligands can be used to target cytokines and other molecules that

50 cooperate synergistically in therapeutic situations in the body of an organism. The description herein thus provides a method for synergizing the activity of two or more cytokines, which comprises the administration of a specific double ligand capable of binding to said two or more cytokines. In this aspect of the description, the double specific ligand can be any double specific ligand, which includes a ligand composed of complementary and / or non-complementary domains, a ligand in an open conformation, and a

55 ligand in a closed conformation. For example, this aspect of the invention relates to combinations of VH and VL domains, VH only domains and VL only domains.

Synergy in a therapeutic context can be achieved in various ways. For example, the target combinations can be therapeutically active only if the ligand is directed to the two targets, while the

Addressing only on a target is not therapeutically effective. In another embodiment, a solo target can provide a low or minimal therapeutic effect, but together with a second target the combination provides a synergistic increase in the therapeutic effect.

Preferably, the cytokines bound by the specific double ligands of this aspect of the description are selected from the list shown in Annex 2.

On the other hand, double specific ligands can be used in oncology applications, in which one specificity is directed to CD89, which is expressed by cytotoxic cells, and the other is tumor specific. Annex 3 provides examples of tumor antigens that can be addressed.

In one embodiment of the second configuration, the variable domains are obtained from an antibody directed against

10 to the first and / or the second antigen or epitope. In a preferred embodiment the variable domains are obtained from a repertoire of individual variable antibody domains. In one example, the repertoire is a repertoire that is not created in an animal or synthetic repertoire. In another example, individual variable domains have not been isolated (at least in part) by animal immunization. Thus, individual domains can be isolated from a new library.

The second configuration, in another aspect, provides a multispecific ligand comprising a first epitope binding domain that has a first epitope binding specificity and a second non-complementary epitope binding domain that has a second epitope binding specificity. . The first and second binding specificities may be the same or different.

In a further aspect, the description provides a multi-specific closed conformation ligand that

20 comprises a first epitope binding domain that has a first epitope binding specificity and a second non-complementary epitope binding domain that has a second epitope binding specificity where the first and second binding specificities are capable of competing by binding to epitopes so that the multispecific ligand of closed conformation cannot bind to the two epitopes simultaneously.

In a further aspect, the description provides open conformation ligands comprising non-complementary binding domains, wherein the domains are specific for a different epitope on the same target. These ligands bind to targets with greater avidity. Similarly, the description provides multivalent ligands comprising non-complementary binding domains specific to the same epitope and targeting targets comprising multiple copies of said epitope, such as IL-5, PDGF-AA, PDGF-BB, TGF-β,

TGF-β2, TGF-β3 and TNFα, for example, as well as human TNF 1 receptor and human TNFα.

In a similar aspect, the ligands according to the invention can be configured to bind to individual epitopes with low affinity, so that binding to individual epitopes is not therapeutically important; however, the increased avidity derived from the binding to two epitopes provides a therapeutic benefit. In a particular example, epitopes that are individually present in normal cell types can be addressed, but they are present together only in abnormal or diseased cells, such as tumor cells. In this situation, only the anomalous or diseased cells are effectively targeted by the bispecific ligands described herein. Specific ligands for multiple copies of the same epitope, or adjacent epitopes, on the same target (known as chelation dAb) may also be trimeric or polymeric ligands (tetrameric or more) comprising three, four or more non-complementary binding domains. By

For example, ligands can be constructed to comprise three or four VH domains or VL domains.

On the other hand, ligands that bind to multisubunit targets are provided, wherein each binding domain is specific for a subunit of said target. The ligand can be dimeric, trimeric or polymeric.

Preferably, multispecific ligands according to the above aspects of the description can be obtained by the method described herein.

According to the previous aspect of the second configuration, advantageously the first epitope binding domain and the second epitope binding domains are variable domains of non-complementary immunoglobulins, as defined herein. That is, they are variable domains vH-vH or vL-vL.

Chelation dAb in particular can be prepared according to a preferred aspect of the description, in particular the use of anchor dAb, in which a dimer, trimeric or multimeric dAb library is constructed using

50 a vector comprising a constant dAb in the 5 ’or 3’ direction of a linking group sequence, with a repertoire of seconds, third parties and more dAb inserted on the other side of the linking group. For example, the anchor or guided dAb can be TAR1-5 (VK), TAR1-27 (VK), TAR2h-5 (VH) or TAR2h-6 (VK).

In alternative methodologies, the use of linker groups can be avoided, for example by the use of non-covalent binding or natural affinity between binding domains such as VH and VK. This document describes a

55 method for preparing a multimeric chelation ligand comprising the steps of:

(to)

provision of a vector comprising a nucleic acid sequence encoding a specific individual binding domain for a first epitope on a target;

(b)

supply of a vector encoding a repertoire comprising second specific binding domains for

a second epitope on said target, so that epitope may be the same as the first epitope or 5 different, said second epitope being adjacent to said first epitope; Y

(C)

expression of said first and second binding domains; Y

(d)

isolation of said combinations of first and second binding domains that are combined together to produce a target binding dimer.

The first and second epitopes are adjacent so that a multimeric ligand is capable of binding to the two

10 epitopes simultaneously. This fact gives the ligand the advantages of greater avidity if it binds. When the epitopes are the same, the increase in avidity is obtained by the presence of multiple copies of the epitope on the target, allowing at least two copies to be joined simultaneously in order to obtain the increase in the avidity effect.

Binding domains can be associated by several methods, as well as by the use of linker groups. By

For example, the binding domains may comprise cis residues, avidin and streptavidin groups or other means for non-covalent fixation after synthesis; These combinations that bind to the target efficiently will be isolated. Alternatively, a linker group may be present between the first and second binding domains, which are expressed as a single polypeptide from a single vector, comprising the first binding domain, the linker group and a repertoire of second binding domains. , for example as described

20 previously.

In a preferred aspect, the first and second binding domains naturally associate when they bind to an antigen; for example, the VH and VL domains (eg VI \), when bound to adjacent epitopes, will naturally associate in a three-way interaction to form a stable dimer. Such associated proteins can be isolated in a target binding assay. An advantage of this procedure is that only the binding domains that

25 bind to closely adjacent epitopes, in the correct conformation, they will associate and thus be isolated as a result of their greater avidity for the target.

In an alternative embodiment of the above aspect of the second configuration of the description, at least one epitope binding domain comprises a "protein framework" or "protein skeleton" not of immunoglobulins as defined herein. Non-immunoglobulin protein frameworks

30 include but preferably are not limited to any of those selected from the group consisting of: SpA, fibronectin, GroEL and other chaperones, lipocalin, CCTLA4 and affibody molecules, as indicated above.

According to the previous aspect of the second configuration of the description, advantageously, the epitope binding domains are fixed to a 'protein skeleton'. Advantageously, a protein skeleton is an immunoglobulin skeleton.

According to the present description, the term "immunoglobulin skeleton" refers to a protein that comprises at least one immunoglobulin folding and acts as a nucleus for one or more epitope binding domains, as defined herein.

Preferred immunoglobulin skeletons as defined herein include any one or more of those selected from the following: an immunoglobulin molecule comprising at least (i) the CL domain (kappa or lambda subclass) of an antibody; or (ii) the CH1 domain of an antibody heavy chain; an immunoglobulin molecule comprising the CH1 and CH2 domains of an antibody heavy chain; an immunoglobulin molecule comprising the CH1, CH2 and CH3 domains of an antibody heavy chain; or any of the subset (ii) in conjunction with the CL domain (subclass kappa or lambda) of an antibody. A hinge region domain can also be included. These domain combinations

45 may, for example, emulate natural antibodies, such as IgG or IgM, or fragments thereof, such as Fv, scFv, Fab or F (ab ’) 2 molecules. Those skilled in the art will appreciate that this list is not intended to be exhaustive.

The binding of the skeleton with the epitope binding domains, as defined herein, can be achieved at the level of polypeptides, that is, after expression of the nucleic acid encoding the skeleton and / or the binding domains to epitopes Alternatively, the linking step can be performed at the level of the

50 nucleic acid Methods for linking a protein skeleton with the one or more epitope binding domains include the use of protein chemistry and / or molecular biology techniques that will be known to those skilled in the art and described herein.

Advantageously, the multispecific closed conformation ligand can comprise a first domain capable of binding to a target molecule, and a second domain capable of binding to a molecule or group that prolongs the half-life of the ligand. For example, the molecule or group may be a bulking agent, such as HSA or a cell matrix protein. As used herein, the phrase "molecule or group that prolongs the

half-life of a ligand "refers to a molecule or chemical group that, when bound by a specific double ligand as described herein, increases the in vivo half-life of said double specific ligand when administered to an animal, with respect to a ligand that does not bind to that molecule or group. Examples of molecules or groups that prolong the half-life of a ligand are described herein below.

5 memory In a preferred embodiment, the multi-specific closed conformation ligand may be able to bind to the target molecule only in the displacement of the molecule or group that enhances the half-life. Thus, for example, a multispecific ligand of closed conformation is kept in circulation in the bloodstream of a subject by means of a volume-forming molecule such as HSA. When a target molecule is found, competition between the binding domains of the multispecific closed conformation ligand causes the displacement of the HSA and the binding of the target.

A ligand according to any aspect of the present description includes a ligand that has or consists of at least one individual variable domain, in the form of an individual variable domain of monomers or in the form of multiple individual variable domains, that is, a multimer. The ligand can be modified to contain additional moieties, such as a fusion protein, or a conjugate. Said multimeric ligand, e.g. eg, in the form of a specific double ligand 15, and / or said ligand comprising or consisting of an individual variable domain, that is, a dAb monomer useful for constructing said multimeric ligand, can advantageously dissociate from its or similar targets. with a Kd of 300 nM or less, 300 nM at 5 pM (i.e., 3 x 10-7 to 5 x 10-12 M), preferably 50 nM to 20 pM, or 5 nM to 200 pM or 1 nM to 100 pM, 1 x 10-7 M or less, 1 x 10-8 M or less, 1 x 10-9 M or less, 1 x 10-10 M or less, 1 x 10-11 M or less; and / or a Koff velocity constant of 5 x 10-1 to 1 x 10-7 S-1, preferably 1 x 10-2 to 1 x 10-6 S-1 or 5 x 10-3 to 1 x 10- 5 S-1 or 5 x 10-1 S-1 or 1 x 10-2 S-1 or less than 1 x 10-3 S-1 or 1 x 10-4 S-1 or 1 x 10-5 S-1 or less or 1 x 10-6 S-1 or as determined, for example, by resonance by surface plasmons. The speed constant Kd is defined as Koff / Kon. In general, a Kd value greater than 1 Molar is considered to indicate non-specific binding. Preferably, an individual variable domain will specifically bind to a target antigen or epitope with a

25 affinity of less than 500 nM, preferably less than 200 nM, and more preferably less than 10 nM, for example less than 500 pM

The description herein provides an anti-TNFα dAb monomer ligand (or double specific ligand comprising said dAb), homodimer, heterodimer or homotrimer, wherein each dAb binds to TNFα. The ligand is bound to TNFα with a Kd of 300 nM at 5 pM (ie, 3 x 10-7 to 5 x 10-12 M), preferably 50 nM to 20 pM, more preferably 5 nM to 200 pM and most preferably from 1 nM to 100 pM; expressed in an alternative way, the Kd is 1 x 10-7 M or less, preferably 1 x 10-8 M or less, more preferably 1 x 10-9 M or less, advantageously 1 x 10-10 M or less and the more preferably 1 x 10-11 M or less; and / or a Koff velocity constant of 5 x 10-1 to 1 x 10-7 S-1, preferably 1 x 10-2 to 1 x 10-6 S-1, more preferably 5 x 10-3 a 1 x 10-5 S-1, for example 5 x 10-1 S-1 or less, preferably 1 x 10-2 S-1 or less, more preferably 1 x

35 10-3 S-1 or less, advantageously 1 x 10-4 S-1 or less, more advantageously 1 x 10-5 S-1 or less, and most preferably 1 x 10-6 S-1 or less, as determined by resonance by surface plasmons.

Preferably, the ligand neutralizes TNFα in a standard L929 assay with DN50 of 500 nM to 50 pM, preferably or 100 nM to 50 pM, advantageously 10 nM to 100 pM, more preferably 1 nM to 100 pM; for example 50 nM or less, preferably 5 nM or less, advantageously 500 pM or less, more preferably 200 pM or less and most preferably 100 pM or less.

Preferably, the ligand inhibits the binding of TNF alpha to TNF alpha I receptor (p55 receptor) with IC50 of 500 nM to 50 pM, preferably 100 nM to 50 pM, more preferably 10 nM to 100 pM, advantageously 1 nM to 100 pM; for example 50 nM or less, preferably 5 nM or less, more preferably 500 pM or less, advantageously 200 pM or less, and most preferably 100 pM or less. Preferably, the TNFα is

45 Human TNFα.

In addition, the description herein provides an anti-TNF I receptor dAb monomer, or double specific ligand comprising said dAb, which is bound to TNF I receptor with a Kd of 300 nM at 5 pM (ie, of 3 x 10-7 to 5 x 10-12M), preferably 50 nM to 20 pM, more preferably 5 nM to 200 pM and most preferably 1 nM to 100 pM, for example 1 x 10-7 M or less, preferably 1 x 10-8 M or less, more preferably 1 x 10-9 M or less, advantageously 1 x 10-10 M or less and most preferably 1 x 10-11 M or less; and / or a Koff velocity constant of 5 x 10-1 to 1 x 10-7 S-1, preferably 1 x 10-2 to 1 x 10-6 S-1, more preferably 5 x 10-3 a 1 x 10-5 S-1, for example 5 x 10-1 S-1 or less, preferably 1 x 10-2 S-1 or less, advantageously 1 x 10-3 S-1 or less, more preferably 1 x 10-4 S-1 or less, more preferably

S-1 S-1

still 1 x 10-5 or less, and most preferably 1 x 10-6 or less, preferably as determined by resonance by surface plasmons.

Preferably, the dAb monomer ligand neutralizes TNFα in a standard assay (e.g., the L929 or HeLa assays described herein) with DN50 of 500 nM to 50 pM, preferably 100 nM to 50 pM, more preferably d 10 nM to 100 pM, advantageously from 1 nM to 100 pM; for example 50 nM or less, preferably 5 nM or less, more preferably 500 pM or less, advantageously 200 pM or less, and most preferably 100 pM or less.

Preferably, the dAb ligand or monomer inhibits the binding of TNF alpha with TNF alpha I receptor (p55 receptor) with an IC50 of 500 nM to 50 pM, preferably 100 nM to 50 pM, more preferably 10 nM to 100 pM, advantageously from 1 nM to 100 pM; for example 50 nM or less, preferably 5 nM or less, more preferably 500 pM or less, advantageously 200 pM or less, and most preferably 100 pM or less.

5 Preferably, the target TNF I receptor is human TNFα.

In addition, the description herein provides a dAb monomer (or specific double ligand comprising said dAb) that is bound to serum albumin (SA) with a Kd of 1 nM at 500 M (i.e., 1 x 10-9 at 5 x 10-4), preferably from 100 nM to 10 M. Preferably, for a specific double ligand comprising a first anti-SA dAb and a second dAb with another target, affinity (eg Kd and / or Koff measured by resonance by

10 superficial plasmons, p. ex. using Biacore) of the second dAb for its target is 1 to 100,000 times (preferably 100 to 100,000, more preferably 1,000 to 100,000 or 10,000 to 100,000 times) the affinity of the first dAb for SA. For example, the first dAb binds to SA with an affinity of approximately 10 M, while the second dAb binds to its target with an affinity of 100 pM. Preferably, the serum albumin is human serum albumin (HSA).

In one embodiment, the first dAb (or a dAb monomer) binds SA (e.g., HSA) with a Kd of about 50, preferably 70, and more preferably 100, 150 or 200 nM.

On the other hand, the description provides dimers, trimers and polymers of the aforementioned dAb monomers.

The ligands according to the invention, which include dAb monomers, dimers and trimers, may be linked to an Fc region of antibodies, comprising one or both domains of CH2 and CH3, and optionally a

20 hinge region. For example, to prepare such polypeptides, vectors encoding ligands linked as a single nucleotide sequence to an Fc region can be used.

In a preferred embodiment of the second configuration, the epitope binding domains are variable domains of immunoglobulins and are selected from repertoires of V genes from individual domains. In general the repertoire of individual antibody domains occurs on the surface of filamentous bacteriophages. In

A preferred embodiment each individual antibody domain is selected by binding a phage repertoire to an antigen.

The present description further provides a kit comprising at least one multi-specific ligand according to the present invention, which may be an open conformation or closed conformation ligand. The kits may be, for example, diagnostic kits, therapeutic kits, kits for the detection of chemical or biological species, and

30 similar.

In a further aspect of the second configuration, the present invention provides a homogeneous immunoassay utilizing a ligand according to the present invention.

In a further aspect of the second configuration, the present invention provides a composition comprising a multi-specific closed forming ligand, which can be obtained by a described method.

Herein, and a pharmaceutically acceptable carrier, diluent or excipient. On the other hand, the present description provides a method for the treatment of a disease using a "multi-specific closed conformation ligand" or a composition described herein.

In a preferred embodiment, the disease is cancer or an inflammatory disease, e.g. ex. rheumatoid arthritis, asthma

or Crohn's disease.

In a further aspect of the second configuration, the present description provides a method for diagnosis, which includes the diagnosis of a disease using a multispecific closed conformation ligand, or a composition described herein. Thus, in general, the binding of an analyte to a multispecific closed conformation ligand can be used to displace an agent, which leads to the generation of a signal in displacement. For example, the binding of the analyte (second antigen) could displace

An enzyme (first antigen) bound to the antibody that provides the basis for an immunoassay, especially if the enzyme is maintained in the antibody through its active site.

Thus, in a final aspect of the second configuration, the present description provides a method for detecting the presence of a target molecule, comprising:

(a) supply of a multi-specific closed forming ligand bound to an agent, said ligand being

50 specific to the target molecule and the agent, wherein the agent that is bound by the ligand leads to the generation of a detectable signal in displacement from the ligand;

(b)

exposure of the multi-specific closed conformation ligand to the target molecule; Y

(C)

detection of the signal generated as a result of agent displacement.

According to the above aspect, advantageously, the agent is an enzyme, which is inactive when bound by the multi-specific closed forming ligand. Alternatively, the agent may be any one or more selected from the group consisting of the following: the substrate for an enzyme, and a fluorescent, luminescent or chromogenic molecule that is inactive or that is deactivated upon binding by the ligand.

5 Calculations of "homology" or "sequence identity" or "similarity" between two sequences (these expressions are used interchangeably herein) are performed as follows. Sequences are aligned for optimal comparison purposes (e.g., gaps may be introduced in one or both between a first and a second amino acid or nucleic acid sequence for optimal alignment and non-homologous sequences may be dismissed for comparison purposes. ). In a preferred embodiment, the length of a reference sequence

10 aligned for comparison purposes is at least 30%, preferably at least 40%, more preferably at least 50%, even more preferably at least 60%, and even more preferably at least 70%, 80% , 90%, 100% of the length of the reference sequence. The amino acid or nucleotide residues at the amino acid positions or the corresponding nucleotide positions are then compared. When a position in the first sequence is occupied by the same amino acid or nucleotide residue as

15 the corresponding position in the second sequence, then the molecules are identical in that position (as used herein the "homology" amino acids or nucleic acids is equivalent to "identity" of amino acids or nucleic acids). The percentage identity between the two sequences is a function of the number of identical positions shared by the sequences, taking into account the number of holes, and the length of each hole, which must be entered for an optimal alignment of the two sequences.

20 Advantageously, the BLAST algorithm (version 2.0) is used for sequence alignment, with parameters set as default values. The BLAST algorithm is described in detail on the website ("www") of the National Center for Biotechnology Information ("NCBI") of the National Institutes of Health ("NIH") of the US government. ("gov"), in the "/ Blast /" directory, in the "blast_help.html" file. The review parameters are defined as follows, and are advantageously set as defined default parameters.

25 BLAST (Basic Local Alignment Search Tool) is the heuristic revision algorithm used by the blastp, blastn, blastx, tblastn and tblastx programs; These programs assign significance to their findings through the statistical methods of Karlin and Altschul, 1990, Proc. Natl Acad. Sci. USA 87 (6): 2264-8 (see file "blast_help.html", as described above) with some improvements. BLAST programs were adapted for sequence similarity search, for example to identify homologs for a review sequence. The

30 programs in general are not useful for searching motif styles. For an explanation of the basic questions about similarity search in sequence databases, see Altschul et al. (1994).

The five BLAST programs available on the National Center for Biotechnology Information website perform the following tasks:

"blastp" compares a review amino acid sequence with a protein sequence database;

"Blastn" compares a review nucleotide sequence with a nucleotide sequence database;

"blastx" compares the six-frame conceptual translation products of a review nucleotide sequence (the two chains) with a protein sequence database;

"tblastn" compares a protein sequence review with a nucleotide sequence database dynamically translated into the six reading frames (the two chains);

40 "tblastx" compares the translations of six frames of a review nucleotide sequence with the translations of six frames of a nucleotide sequence database.

BLAST uses the following review parameters:

HISTOGRAM Displays a histogram of scores for each search; The default value is yes. (See parameter H in the BLAST Manual).

45 DESCRIPTIONS Limits the number of short descriptions of matched sequences referring to the specified number; The default limit is 100 descriptions. (See parameter V on the manual page). See also EXPECT and CUTOFF.

ALIGNMENTS Limits database sequences to the specified number for which high-score segment pairs (HSP) refer; the default limit is 50. If a larger number of base sequences of

50 data meets the threshold of statistical significance for communication (see EXPECT and CUTOFF below), only the correspondences assigned to the greatest statistical significance are communicated. (See parameter B in the BLAST Manual).

EXPECT The threshold of statistical significance for communicating correspondences with database sequences; the default value is 10, so it is expected that 10 matches are found by

coincidence, according to the stochastic model of Karlin and Altschul (1990). If the statistical significance attached to a correspondence is greater than the EXPECT threshold, the correspondence will not be communicated. The lower EXPECT thresholds are more severe, which leads to the communication of fewer correspondences by chance. Fractional values are acceptable. (See parameter E in the BLAST Manual).

5 CUTOFF Cut score to communicate pairs of high score segments. The default value is calculated from the EXPECT value (see above). HSP values are communicated for a sequence of databases only if the statistical significance attached to them is at least as high as it would be ascribed to a single HSP value that has a score equal to the CUTOFF value. Higher CUTOFF values are more severe, which leads to the communication of fewer correspondences by chance. (See parameter S in the

10 BLAST manual). Normally, significance thresholds can be handled more intuitively through EXPECT.

MATRIX Specifies an alternative score matrix for BLASTP, BLASTX, TBLASTN and TBLASTX. The default matrix is BLOSUM62 (Henikoff & Henikoff, 1992, Proc. Natl. Acad. Sci. USA 89 (22): 10915-9). Valid alternative choices include: PAM40, PAM120, PAM250 and IDENTIFY. There are no scoring matrices

15 alternatives available for BLASTN; to specify the MATRIX directive in BLASTN, a response to the error must be returned.

STRAND Limits the TBLASTN search to the upper or lower string of the database sequences; or limits the BLASTN, BLASTX or TBLASTX search to only the reading frames of the upper or lower chain of the review sequence.

20 FILTER Masks segments of the review sequence that have low computational complexity as determined by the SEG program of Wootton & Federhen (1993) Computers and Chemistry 17: 149-163, or segments consisting of low periodicity internal repeats, determined by the program XNU of Claverie & States, 1993, Computers and Chemistry 17: 191-201, or, for BLASTN, by the DUST program of Tatusov and Lipman (see the NCBI website). Filtering can eliminate statistically significant reports but without biological interest from

25 blast output (e.g., impacts against common hill-rich, acidic or basic regions), which leaves the biologically most interesting regions of the review sequence available for specific correspondence versus database sequences .

The low complexity sequence found by a filter program is substituted using the letter "N" in the nucleotide sequences (eg, "N" repeated 13 times) and the letter "X" in the protein sequences (p eg, "X"

30 repeated 9 times).

Filtering is applied only to the review sequence (or its translation products), not to database sequences. The default filtering is DUST for BLASTN, SEG for other programs.

It is not uncommon that with SEG, XNU or both, nothing is masked when applied to sequences in SWISS-PROT, so it is not expected that filtering will always have an effect. In addition, in some cases, the

35 sequences are completely masked, indicating that the statistical significance of any correspondence communicated against the unfiltered review sequence should be suspected.

NCBI-gi Makes the NCBI gi identifiers displayed at the exit, in addition to the access name and / or location.

It is more preferable that sequence comparisons are made using the simple BLAST review algorithm provided on the NCBI website described above, in the "/ BLAST" directory.

40 Brief description of the figures

Figure 13 VH training sequence for library 1. The VH frame sequence based on the DP47-JH4b germline sequence. The positions in which the NNK randomization (N = nucleotides A or T or C or G; K = nucleotides G or T) has been incorporated in library 1 are indicated in the underlined bold text.

Figure 14 VH training sequence for library 2. The VH frame sequence based on the sequence of

45 germline DP47 -JH4b. The positions in which the NNK randomization (N = nucleotides A or T or C or G; K = nucleotides G or T) has been incorporated into library 2 are indicated in the underlined bold text.

Figure 15 Training Vκ sequence for library 3. The Vκ frame sequence based on the DPK9-JK1 germline sequence. The positions in which the NNK randomization (N = nucleotides A or T or C or G; K = nucleotides G or T) has been incorporated into library 3 are indicated in the underlined bold text.

50 Figure 16 Nucleotide and amino acid sequence of dAb anti MSA MSA 16 and MSA 26.

Figure 17 Biacore inhibition of MSA 16 and 26. Purified dAb MSA16 and MSA26 were analyzed by Biacore inhibition to determine Kd. Briefly, the dAb were tested to determine the concentration of dAb needed to achieve 200 RU of response on a Biacore CM5 chip coated with high MSA density. Once the required concentrations of dAb were determined, MSA antigen was premixed in a range of

concentrations around the expected Kd with dAb and incubated overnight. The binding to the BAcore chip coated with MSA of dAb in each of the premixes with a high flow rate of 30 µl / minute was then measured.

Figure 18 Serum levels of MSA16 after injection. The serum half-life of dAb MSA16 was determined in

5 mouse The MSA16 was dosed as IV injections. individual at approximately 1.5 mg / kg in CD1 mice. Modeling with a 2-compartment model showed that MSA16 had a t1 / 2α value of 0.98 h, a t1 / 2β of 36.5 h and an ABC of 913 h.mg/ml. MSA16 had a considerably elongated half-life compared to HEL4 (a chicken egg white antilisozyme dAb) that had a t1 / 2α of 0.06 h and a t1 / 2β of 0.34 h.

Figure 19 ELISA (a) and TNF receptor assay (c) showing the inhibition of TNF binding with a

10 Fab type fragment comprising MSA26Ck and TAR1-5-19CH. The addition of MSA with the Fab type fragment reduces the level of inhibition. An ELISA plate coated with 1 mg / ml of TNFα was probed with Vκ and Cκ double specific Fab type fragment of Vκ and also with a control TNFα binding dAb at a concentration calculated to provide a similar signal in ELISA. The specific double and control dAb were used to probe the ELISA plate in the presence and absence of 2 mg / ml of MSA. The signal in the specific well was reduced

15 doubles by more than 50% but the signal in the dAb well was not reduced at all (see figure 19a). The same double specific protein was also placed in the receptor assay with and without MSA and competition for MSA was also shown (see Figure 19c). This demonstrates that the binding of MSA with the specific double is competitive with the binding to TNFα.

Figure 20 TNF receptor assay showing inhibition of TNF binding with a dAb disulfide-linked heterodimer of TAR1-5-19 and dAb of MSA16. The addition of MSA with the dimer reduces the level of inhibition 20 in a dose-dependent manner. The TNF receptor assay (Figure 19 (b)) was performed in the presence of a constant heterodimer concentration (18 nM) and a dilution series of MSA and HSA. The presence of HSA in a concentration range (up to 2 mg / ml) did not cause a reduction in the ability of the dimer to inhibit TNFα. However, the addition of MSA caused a dose-dependent reduction in the ability of the dimer to inhibit TNFα (Figure 19a). This demonstrates that MSA and TNFα compete for binding to the MSA16 dimer TAR1-5-19

25 joined in cis. MSA and HSA alone had no effect on the level of TNF binding in the trial.

Figure 21 Purified recombinant human serum albumin (HSA) domains, lanes 1-3 contain HSA domains I, II and III, respectively.

Figure 22 Example of an immunoprecipitation showing that the HSA binding dAB binds to full-length HSA (lane 8) and HSA II domain (lane 6), but does not bind to HSA domains I and III (lanes 5 and 7,

30 respectively). An HSA non-binding dAb does not lower full-length HSA or HSA I, II or III domains (lanes 14).

Figure 23. Example of identification of binding to an HSA domain by a dAb identified by resonance by surface plasmons. The dAb under study was injected as described in a low density coated human serum albumin CM5 sensor chip (Biacore). At point 1, the dAb under study was injected alone at 1 M. 35 At point 2, using the co-injection order, the same injection was changed to a mixture of dAb 1 M plus HSA domain 1, 2 or 3 7 M, produced in Pichia. At point 3, the sample injection was interrupted, and the buffer flow continued. The results for two different dAb are shown in 23 a) and 23 b). When dAb is injected with the HSA domain to which it binds, it forms a complex that can no longer bind to HSA on the chip, and therefore the Biacore signal decreases at point 2, with a loss of speed that reflects the balance between pathways between dAb, HSA domain

40 soluble and chip attached to HSA. When the domain does not bind dAb, the signal remains unchanged at point 2, and begins to descend only at point 3, where the flow is changed to the buffer. In these two cases, dAb binds to the HSA 2 domain.

Figure 24 Antibody sequences of AlbudAb ™ clones (a dAb that specifically binds to serum albumin) identified by phage selection. All clones have aligned in the human germline genes. The

45 residues that are identical to the germ line have been represented by ‘.’. In CDR3 VH, the symbol ‘a-’ has been used to facilitate alignment, but does not represent a residue. All clones were selected from libraries based on a single human framework comprising the V3-23 / DP47 and JH4b heavy chain germline genes for the VH libraries and the κ O12 / O2 / DPK9 and Jκ1 light chain genes for the Vκ libraries with a variety of side chains incorporated into antigen binding site positions.

50 Figure 25 Alignments of the three domains of human serum albumin. The conservation of cysteine residues can be clearly observed.

Detailed description of the invention

Definitions

"Complementary" Two immunoglobulin domains are "complementary" when they belong to families of

55 structures that form pairs or similar groups or are obtained from these families and retain this characteristic. For example, a VH domain and a VL domain of an antibody are complementary; two VH domains are not complementary and two VL domains are not complementary. Complementary domains can be found in

other members of the immunoglobulin superfamily, such as the Vα and Vβ domains (or γ and δ) of the T lymphocyte receptor. In the context of the second configuration of the present invention, the non-complementary domains do not cooperatively bind a target molecule. , but act independently on different target epitopes that may be in the same or in different molecules. Domains that are artificial, such as domains

5 based on protein frameworks that do not bind epitopes unless they are designed for this purpose, are not complementary. Similarly, two domains based (for example) on an immunoglobulin domain and a fibronectin domain are not complementary.

"Immunoglobulin" Refers to a family of polypeptides that retain the immunoglobulin folding characteristic of antibody molecules, which contains two β-sheets and, normally, a conserved disulfide bond. Members of the immunoglobulin superfamily are involved in many aspects of cellular and non-cellular interactions in vivo, which includes extensive functions in the immune system (e.g., antibodies, T-cell receptor molecules and the like), involvement in the cell adhesion (for example ICAM molecules) and intracellular signaling (for example, receptor molecules, such as the PDGF receptor). The present invention is applicable to all molecules of the immunoglobulin superfamily that

15 have binding domains. Preferably, the present invention relates to antibodies.

"Combination" The variable domains according to the invention combine to form a group of domains; for example, complementary domains can be combined, for example by combining VL domains with VH domains. Non-complementary domains can also be combined. The domains can be combined in various ways, which involve the binding of the domains by covalent or non-covalent means.

20 "Domain" A domain is a folded protein structure that retains its tertiary structure independently of the rest of the protein. In general, the domains are responsible for the discrete functional properties of the proteins, and in many cases they can be added, removed or transferred to other proteins without loss of function of the rest of the protein and / or the domain.

As used herein, an "individual variable domain" is a domain that can be linked.

Specifically to an epitope, an antigen or a ligand independently, that is, without the need for another binding domain to cooperatively bind to the epitope, antigen or ligand. Said epitope, antigen or ligand may be of natural occurrence, or it may be a modification of an epitope, antigen or ligand of natural occurrence, or it may be synthetic. The "variable" part of the individual variable domain essentially determines the binding specificity of each individual variable domain in particular. Thus, the term "variable" in the context of domains

30 individual variables, refers to the fact that sequence variability is not uniformly distributed across an individual variable domain, but is essentially distributed between the frame or skeleton parts of the individual variable domain. For example, in an individual variable domain of antibodies, the variability is concentrated in one to three segments commonly known as complementarity determination regions (CDR). The one or more CDRs can be distributed among the main structure regions of the

Antibody (FR) of a light chain or heavy chain to form an individual variable domain of light chain of antibodies or an individual variable domain of heavy chain of antibodies, respectively, each of which specifically binds to an epitope independently from another binding domain. With an analogous structure, there is an individual variable domain of T-lymphocyte receptor, with its one to three CDRs distributed between the TCR framework domains.

Thus, the variable parts that confer the binding specificity of individual variable domains may differ widely in sequence from other individual variable domains that have substantially the same part of the remaining framework, and therefore may have a range of different binding specificities. Individual variable domain frameworks include frameworks of the principal structure of the antibody, principal structures of consensus antibodies and frameworks that originate and / or originate from bacterial proteins, e.g. ex.

45 GroEL, GroEs, SpA, SpG, and eukaryotic proteins, e.g. eg, CTLA-4, lipocalins, fibronectin, etc. A source of the variable portions of individual variable domains includes one or more CDRs, which can be grafted onto non-immunoglobulin frameworks as well as frameworks of principal antibody structures to generate individual variable domains of antibodies. Another source of variation in an individual variable domain may be the diversification of positions chosen in a framework framework not of immunoglobulins such as fibronectin, to generate domains

50 individual variables, using molecular biology techniques, such as NNK codon diversity. Similarly, this source of variation is also applicable to an individual variable domain of antibodies.

An individual variable antibody domain can be obtained from antibody sequences encoded and / or generated by an antibody producing species, and includes fragment or fragments and / or derivatives of the antibody variable region, including one or more framework regions, or sequences of framework consensus, and / or one or more 55 CDR. Accordingly, an individual variable domain of antibodies includes fragments and / or derivatives of a light chain variable region of antibodies, or a heavy chain variable region of antibodies, or a region of VHH antibodies. For example, regions of VHH antibodies include those that are endogenous to camelids: p. eg, camels and llamas, and the new antigen receptor (NAR) of nurse sharks or orectolóbidos (Roux et al., 1998 PNAS 95 (20): 11804-9) and the VH region of spotted ratfish (Rast et al ., 1998 60 Immunogenetics 47: 234-245). Variable domains of antibody light chain and variable domains of antibody heavy chain include the endogenous of an animal species that includes, but preferably not

is limited to, human being, mouse, rat, pig, cynomolgus, hamster, horse, cow, goat, dog, cat and birds, p. ex. Human VKappa and human VH3, respectively. The variable regions of the light chain of antibodies and the variable regions of the heavy chain of antibodies also include a main structure of consensus antibodies, as described below, which includes the families of the V region, such as the VH3 family. A domain

5 individual variable T cell receptor is an individual variable domain that comes from a chain or chains of T cell receptor, e.g. eg, α, β, γ and δ chains, and that bind to an epitope or an antigen or a ligand independently of another binding domain for that epitope, antigen or ligand, analogously to the individual variable domains of antibodies.

An individual variable domain of antibodies also comprises a protein domain comprising a

10 framework that is not derived from an antibody or T lymphocyte receptor, and that has been genetically engineered to show diversity in binding specificity with respect to a state prior to design, incorporating into the framework, one or more of a CDR1 , a CDR2 and / or a CDR3, a derivative or a fragment thereof, or a complete V antibody domain. An individual variable domain of antibodies may also include a non-immunoglobulin framework and immunoglobulin frameworks as illustrated in the variable domain multimers.

15 individual GroEL described below. Preferably the CDR (s) come from an antibody domain V of an antibody chain, e.g. eg, VH, VL and VHH. The antibody chain can specifically bind an antigen or epitope in concert with a second antibody chain, or the antibody chain can specifically bind an antigen or epitope independently of a second antibody chain, such as a VHH chain. The integration of one or more CDRs into an individual variable domain of antibodies that

20 comprises a non-immunoglobulin framework must derive in the ability of the individual variable domain of the non-immunoglobulin framework to specifically bind an epitope or an antigen or a ligand independently of another binding domain for that epitope, antigen or ligand.

An individual domain is transformed into an individual variable domain by introducing diversity in the site (s) designed to become binding sites, followed by the selection of desired binding characteristics 25 using, for example, presentation technologies. Diversity can be introduced at specific sites of a non-immunoglobulin framework of interest by randomizing the amino acid sequence of specific framework bonds, e.g. ex. introducing NNK codons. This mechanism for generating diversity followed by the selection of desired binding characteristics is similar to the natural selection of antibodies specific for high affinity antigens that come from the diversity generated in the bonds that make up the antibody binding site of

30 nature. Ideally, an individual domain that is small and contains a folding similar to that of an antibody loop, is transformed into an individual variable domain, variants of the individual variable domain are expressed from which individual variable domains containing specificities and characteristics can be selected desired binding between libraries containing a large number of variants of the individual variable domain.

35 Nomenclature of the individual variable domains: sometimes the nomenclature of an individual variable domain of antibodies is abbreviated by removing the first "d" or the letters "Dom", for example, Ab7h24 is identical to dAb7h24 which is identical to DOM7h24.

By individual variable domain of antibodies is meant a folded polypeptide domain comprising characteristic sequences of variable antibody domains. Therefore it includes variable antibody domains

Complete and modified variable domains, for example, in which one or more bonds have been replaced by sequences that are not characteristic of variable antibody domains, or variable antibody domains that have been truncated or comprise extensions at the N or C end , as well as folded fragments of variable domains that retain at least in part the union activity and specificity of the full length domain.

"Repertoire" Collection of various variants, for example polypeptide variants that differ in sequence

45 primary. A library used in the present invention will comprise a repertoire of polypeptides comprising at least 1,000 members.

"Library" The term library refers to a mixture of heterogeneous polypeptides or nucleic acids. The library is composed of members, each of which has a unique sequence of polypeptides or nucleic acids. In this sense, library is synonymous with repertoire. Sequence differences between the 50 members of a library are responsible for the diversity present in the library. The library may take the form of a simple mixture of polypeptides or nucleic acids, or it may be in the form of organisms or cells, for example bacteria, viruses, animal or plant cells and the like, transformed with a nucleic acid library. Preferably, each individual organism or cell contains only one or a limited number of library members. Advantageously, the nucleic acids are incorporated into expression vectors, in order to allow the expression of the polypeptides encoded by the nucleic acids. In a preferred aspect, therefore, a library can take the form of a population of host organisms, each organism containing one or more copies of an expression vector containing a single member of the library in the form of a nucleic acid that can be expressed for produce its corresponding polypeptide member. Thus, the population of host organisms has the possibility of encoding a large repertoire of genetically diverse polypeptide variants.

A "multispecific closed conformation ligand" describes a multispecific ligand as defined in the

present specification comprising at least two epitope binding domains as defined herein. The term "closed conformation" (multispecific ligand) means that the epitope binding domains of the ligand are arranged such that epitope binding by an epitope binding domain competes with epitope binding by another epitope binding domain. That is, similar epitopes can

5 be bound by each epitope binding domain individually but not simultaneously. The closed conformation of the ligand can be achieved using the methods described herein.

"Antibody" An antibody (for example IgG, IgM, IgA, IgD or IgE) or fragment (such as Fab, F (ab ') 2, Fv, Fv with disulfide bond, scFv, specific closed conformation antibody, scFv with bond of disulfide, diabody) either from any species that naturally produces an antibody or is believed by DNA technology

10 recombinant; isolated from serum, B lymphocytes, hybridomas, transfectomas, yeasts or bacteria).

"Double specific ligand" A ligand comprising a first individual variable domain of immunoglobulins and a second individual variable domain of immunoglobulins as defined herein, wherein the variable domains are capable of binding to two different antigens or two epitopes in the same antigen that are not normally bound by a monospecific immunoglobulin. For example, the two epitopes may be in the same hapten, but they are not in the same epitope or adjacent enough to be linked by a monospecific ligand. The specific double ligands according to the invention are composed of variable domains that have different specificities, and do not mutually contain pairs of complementary variable domains that have the same specificity. Thus, the specific double ligands, which as defined herein, contain two individual variable domains, are a subset of multimeric ligands, which as defined herein contain two or more individual variable domains, wherein at least Two of the individual variable domains are capable of binding to two different antigens or to two different epitopes on the same antigen. In addition, a specific double ligand as defined herein is also distinct from a ligand comprising an individual variable domain of antibodies, and a second antigen and / or epitope binding domain that is not an individual variable domain. Likewise, a specific double ligand

25 as defined herein is also distinct from a ligand containing a first and second epitope binding antigen / domain, in which the epitope binding antigen / domain is not an individual variable domain as defined in This memory.

"Antigen" A molecule that is bound by a ligand according to the present invention. Normally, the antigens are bound by antibody ligands and are capable of raising the response of an antibody in vivo. May

30 it is a polypeptide, protein, nucleic acid or other molecule. In general, the specific double ligands according to the invention are selected for their specificity with respect to the target against a given antigen. In the case of conventional antibodies and fragments thereof, the antibody binding site defined by the variable bonds (L1, L2, L3 and H1, H2, H3) can bind to the antigen.

"Epitope" A unit of structure conventionally linked by a pair of VH / VL immunoglobulins. Epitopes

35 define the minimum binding site for an antibody, and thus represent the specificity target of an antibody. In the case of an individual domain antibody, an epitope represents the unit of structure linked by an isolated variable domain. An epitope binding domain comprises a protein framework and epitope interaction sites (which are advantageously on the surface of the protein framework). An epitope binding domain may comprise epitope interaction sites that are nonlinear, e.g. ex. in which the binding domain to

40 epitopes comprise multiple interaction sites of epitopes that have intermediate regions between them, e.g. eg, CDR separated by FR, or are present in separate polypeptide chains. Alternatively, an epitope binding domain may comprise a linear epitope interaction site composed of continuously encoded amino acids in a polypeptide chain.

"Generic ligand" A ligand that is linked to all members of a repertoire. In general, it doesn't bind through

45 of the antigen binding site as defined above. Some non-limiting examples include protein A, protein L and protein G.

"Selection" Obtained by screening, or from a Darwinian selection process, in which binding interactions take place between a domain and the antigen or epitope or between an antibody and an antigen or epitope. Thus a first variable domain can be selected to bind an antigen or epitope in the presence or absence of a

50 complementary variable domain.

"Universal framework" Main structure sequence of the individual antibody corresponding to the regions of an antibody conserved in sequence as defined in Kabat ("Sequences of Proteins of Immunological Interest", US Department of Health and Human Services) or corresponding to the repertoire or structure of human germline immunoglobulins as defined in Chothia and Lesk, (1987) J. Mol. Biol. 196: 910-917. The invention

55 provides the use of a single frame, or a set of said frames, which as it has been discovered allows to obtain virtually any specificity of binding through variation in the hypervariable regions alone.

As used herein, "conjugate" refers to a composition comprising an antigen-binding fragment of an antibody that binds to serum albumin that is bound to a drug.

As used herein, the term "small molecule" means a compound having a molecular weight of less than 1,500 dalton, preferably less than 1,000 dalton.

Such conjugates include "drug conjugates", which comprise an antigen-binding fragment of an antibody that binds to serum albumin to which a drug is covalently bound, and "non-covalent drug conjugates", which comprise a fragment of Antigen binding of an antibody that binds to serum albumin to which a drug is not covalently bound.

As used herein, "drug conjugate" refers to a composition comprising an antigen-binding fragment of an antibody that binds to serum albumin to which a drug is covalently bound. The drug may be covalently bound to the direct antigen binding fragment or

10 indirectly through a suitable linker group moiety. The drug can be attached to the antigen-binding fragment in any suitable position, such as the amino end, the carboxyl end or through suitable amino acid side chains (eg, the lysine amino group).

"Half-life" Time taken for the serum concentration of the ligand to be reduced to 50%, in vivo, for example due to degradation of the ligand and / or clearance or sequestration of the ligand by natural mechanisms. The ligands of the invention are stabilized in vivo and their half-life is increased by binding to molecules that resist degradation and / or clearance or sequestration. Normally, these molecules are naturally occurring proteins that have a long half-life in vivo. The half-life of a ligand increases if its functional activity persists, in vivo, for a longer period than a similar ligand that is not specific to the molecule that increases the half-life. Thus, a specific ligand for HSA and a target molecule is compared with the same ligand in which

20 the specificity for HSA is not present, which does not bind HSA but binds to another molecule. For example, it can bind to a second epitope in the target molecule. Normally, the half-life increases by 10%, 20%, 30%, 40%, 50% or more. Increases in the range of 2x, 3x, 4x, 5x, 10x, 20x, 30x, 40x, 50x or more of the half-life are possible. Alternatively, or in addition, increases in the range of up to 30x, 40x, 50x, 60x, 70x, 80x, 90x, 100x, 150x of the half-life are possible.

The phrase "substantially the same" when used to compare the beta T half-life of a ligand with the beta T half-life of serum albumin in a host means that the beta T half-life of the ligand in a host varies no more than 50% with respect to to the beta T half-life of serum albumin itself in the same host, preferably a human host, e.g. For example, the beta T half-life of said ligand is no more than 50% less or no more than 50% greater than the beta T half-life of serum albumin in a specified host. Preferably,

30 when referring to the phrase "substantially the same", the beta T half-life of the ligand in a host varies no more than 20% to 10% with respect to the serum albumin half-life itself, and more preferably, varies no more of 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2% or 1%, at least with respect to the serum albumin half-life itself, or not Nothing varies with respect to the serum albumin half-life itself.

Alternatively, the phrase "not substantially the same" when used to compare the beta T half-life of a

Ligand with the serum albumin beta T half-life in a host means that the beta T half-life of the ligand in a host varies at least 50% with respect to the serum-albumin beta T half-life itself in the same host, preferably a host human, p. For example, the beta T half-life of the ligand is more than 50% greater than the beta T half-life of serum albumin in a specified host.

"Homogeneous immunoassay" An immunoassay in which analyte is detected without the need for a step of separation of bound or unbound reagents.

"Substantially identical" or "substantially homologous" A first amino acid or nucleotide sequence containing a sufficient number of identical or equivalent amino acid residues or nucleotides (eg, with a similar side chain, eg amino acid substitutions conserved) to a second amino acid or nucleotide sequence so that the first or second amino acid or nucleotide sequences have activities

45 similar. In the case of first and second antibodies and / or individual variable domains described herein, the second antibody or individual variable domain has the same binding specificity as the first and has at least 50%, or at least up to 55 %, 60%, 70%, 75%, 80%, 90%, 95%, 96%, 97%, 98%, 99% or 100% of the affinity of the first antibody or individual variable domain.

As used herein, the terms "conditions of low severity", "medium severity",

50 "high severity", or "very high severity" describe conditions for hybridization and washing of nucleic acids. Orientation for hybridization reactions can be found in Current Protocols in Molecular Biology, John Wiley & Sons, NY (1989), 6.3. 1-6.3.6, which is incorporated herein by reference in its entirety Aqueous and non-aqueous methods are described in that reference and which can be used.The specific hybridization conditions referred to herein are as follows: (1 ) hybridization conditions of

55 low severity in sodium chloride / sodium citrate (SSC) 6X at approximately 45 ° C, followed by two washes in 0.2X SSC, 0.1% SDS at least 50 ° C (wash temperatures may increase to 55 ° C for conditions of low severity); (2) hybridization conditions of medium severity in 6X SSC at approximately 45 ° C, followed by one or more washes in 0.2X SSC, 0.1% SDS at 60 ° C; (3) high severity hybridization conditions in 6X SSC at approximately 45 ° C, followed by one or more washes in 0.2X SSC, 0.1% SDS a

65 ° C; and preferably (4) hybridization conditions of very high severity which are 0.5 M sodium phosphate, 7% SDS at 65 ° C, followed by one or more washes at 0.2X SSC, 1% SDS at 65 ° C . Conditions of very high severity (4) are the preferred conditions and those that should be used unless otherwise specified.

"Resonance by superficial plasmons" Competition tests can be used to determine if a

5 specific antigen or epitope, such as human serum albumin, competes with another antigen or epitope, such as cynomolgus serum albumin, for binding to a serum albumin binding ligand described herein, as a specific dAb. Similarly, competition assays can be used to determine whether a first ligand such as dAb competes with a second ligand as a dAb for binding to a target antigen or epitope. The term "competes" as used herein refers to a substance, such as a molecule, compound,

Preferably a protein, which is capable of interfering to some extent with the specific binding interaction between two or more molecules. The phrase "does not competitively inhibit" means that the substance, such as a molecule, a compound, preferably a protein, does not interfere in any measurable or significant extent with the specific binding interaction between two or more molecules. The specific binding interaction between two or more molecules preferably includes the specific binding interaction between an individual variable domain and its

15 participant or similar target. The interference or competition molecule may be another individual variable domain or it may be a molecule that is structurally and / or functionally similar to a similar participant or target.

An individual variable domain includes an individual variable domain of immunoglobulins and a single non-immunoglobulin variable domain that contains one, two, three or more CDR regions of an immunoglobulin variable domain, such as an antibody variable domain, which includes an individual variable domain chain

20 heavy or light chain antibody. The individual variable domain can be obtained from an animal, which includes a human being, rat, mouse, pig, monkey, camelid, as a variable region (V) of antibodies, or can be obtained from a microorganism such as E. coli in the case of non-immunoglobulin framework of GroEL and GroEs. An individual variable domain can be partially or totally artificial, or it can be generated using recombinant molecular biology technology.

25 In vitro proficiency tests to determine the ability of an individual variable domain to compete for binding to a target in another target binding domain, as another individual variable domain, as well as to determine Kd, are well known in the technique. A preferred competition test is a surface plasmon resonance test, which has the advantages of being fast, sensitive and useful over a wide range of protein concentrations, and which requires small amounts of sample material. An essay of

30 preferred surface plasmon resonance competition is a Biacore competition experiment. A Biacore competition experiment can be used to determine whether, for example, cynomolgus serum albumin and human serum albumin compete for binding to a ligand such as dAb DOM7h-x. An experimental protocol for this example is as follows.

For example, after coating a CM5 sensor chip (Biacore AB) at 25 ° C with approximately 1,000

35 resonance units (RU) of human serum albumin (HSA), a purified dAb is injected onto the surface of the antigen at a single concentration (e.g., 1 µm) alone, and in combination with a dilution series of cynomolgus serum albumin (CSA). Serial dilutions of HSA were mixed with a constant concentration (40 nM) of the purified dAb. A suitable dilution series of CSA would be to start with 5 M CSA, with six double dilutions up to 78 nM CSA. Let these solutions reach equilibrium before

40 of the injection. After injection, a response reading is taken to measure the resulting binding RUs for the dAb alone and each of the various dAb / CSA mixtures, using the data according to the BIA evaluation software, generates a curve of dose-response for each CSA inhibition of AlbudAb ™ binding (a dAb that specifically binds to serum albumin) to the chip in which the HSA is immobilized. By comparing the united RU of dAb alone with the united RU of dAb + CSA, it will be possible to see if the CSA competes with the HSA to join

45 to dAb. If there is competition, then when the concentration of CSA in solution increases, the RUs of dAb bound to HSA will decrease. If there is no competition, the addition of CSA will have no impact on the amount of dAb that is bound to HSA.

An expert will know how to adapt this or other protocols to perform this competition assay on several different ligands, which includes the different ligands described herein that bind to serum albumin. The

The variety of ligands includes, but is not limited to, the individual variable domains of monomers, which include individual variable domains comprising a framework of immunoglobulins and / or non-immunoglobulins, dAb, specific double ligands and multimers of these ligands. An expert would also know how to adapt this protocol to compare the binding of several pairs of different antigens and / or epitopes to a ligand using this competition assay.

These competition experiments can provide a numerical cut-off value according to which it can be determined whether an antigen or epitope competes with another antigen or epitope for binding to a specific ligand, preferably dAb. For example, in the experiment described above, if 5 M CSA in solution produces a reduction of 10%, or less, reduction in RU of binding of dAb to HSA, then it is considered that there is no competition for binding. Consequently, a reduction in the binding RUs of dAb to HSA in the presence of CSA

60 of more than 10% would indicate the presence of competition for the binding of dAb for binding to HSA by CSA. A reduction in RU binding of dAb to HSA of less than 10% would indicate the absence of competition for CSA for

the binding of dAb to HSA, with reductions of 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2% and 1% which are progressively more severe requirements to indicate the absence of competition. The greater the reduction in the binding RUs of dAb to HSA, the higher the competition. Thus, increasing levels of competition can be graded according to the percentage reduction in the HSA-binding RUs, that is, at least 15%, 20%, 25%, the

5 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90 %, 95% or up to 100% reduction.

A fragment as used herein refers to less than 100% of the sequence (e.g., up to 99%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, 10%, etc.)., but comprising 5, 6, 7, 8, 9, 10, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 or more contiguous amino acids. A fragment is of sufficient length such that the binding to the serum albumin of interest is maintained with an affinity of 1 x 10-6 M or less. A fragment as used herein also refers to insertions, deletions and optional substitutions of one or more amino acids that do not substantially alter the ability of the altered polypeptide to bind to an individual domain antibody obtained with respect to the target. The number of amino acid insertions, deletions or substitutions is preferably up to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,

15 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 , 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65 , 66, 67, 68, 69 or 70 amino acids.

Detailed description of the invention

Unless stated otherwise, all the technical and scientific terms used herein have the same meaning as is normally understood by one skilled in the art (eg, in cell culture, molecular genetics, nucleic acid chemistry, techniques of hybridization and biochemistry). Standard techniques are used for molecular, genetic and biochemical methods (see in general, Sambrook et al., Molecular Cloning: A Laboratory Manual, 2nd ed. (1989) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY and Ausubel et al. , Short Protocols in Molecular Biology (1999) 4th Ed, John Wiley & Sons, Inc. which are incorporated herein by reference) and chemical methods. Standard techniques for plasmon resonance assays

25 superficial include Jan Terje Andersen et al. (2006) Eur. J. Immunol. 36: 304-3051 Fagerstam (1991) Tech. Chem Protein. 2: 65-71, and Johnsson et al. (1991) Anal. Biochem 198: 268-277.

Preparation of multispecific ligands based on immunoglobulins

The specific double ligands described herein, whether they have open or closed conformation according to the desired configuration, can be prepared according to previously established techniques, used in the field of antibody design, for the preparation of scFv, "phage" antibodies and other designed antibody molecules. Techniques for the preparation of antibodies, and in particular bispecific antibodies, are described for example in the following reviews and in the references cited therein: Winter & Milstein, (1991) Nature 349: 293-299; Plueckthun (1992) Immunological Reviews 130: 151-188; Wright et al., (1992) Crit. Rev. Immunol. 12: 125-168; Holliger, P. & Winter, G. (1993) Curr. Op. Biotechn. 4, 446-449; Carter, et al. (1995) J. Hematother. 4,

35 463-470; Chester, K.A. & Hawkins, R.E. (1995) Trends Biotechn. 13, 294-300; Hoogenboom, H.R. (1997) Nature Biotechnol. 15, 125-126; Fearon, D. (1997) Nature Biotechnol. 15, 618-619; Plückthun, A. & Pack, P. (1997) Immunotechnology 3, 83-105; Carter, P. & Merchant, A.M. (1997) Curr. Opin. Biotechnol 8, 449-454; Holliger, P. & Winter, G. (1997) Cancer Immunol. Immunother 45,128-130.

The description herein provides the selection of variable domains against two different antigens or 40 epitopes, and the subsequent combination of the variable domains.

The techniques used for the selection of variable domains employ libraries and selection procedures that are known in the art. Natural libraries (Marks et al. (1991) J. Mol. Biol., 222: 581; Vaughan et al. (1996) Nature Biotech., 14: 309) that use rearranged V genes collected from human B lymphocytes are well known for those skilled in the art. The synthetic libraries (Hoogenboom & Winter (1992) J. Mol.

45 Biol., 227: 381; Barbas et al. (1992) Proc. Natl Acad. Sci. USA, 89: 4457; Nissim et al. (1994) EMBO J., 13: 692; Griffiths et al. (1994) EMBO J., 13: 3245; From Kruif et al. (1995) J. Mol. Biol., 248: 97) are prepared by cloning of immunoglobulin V genes, usually using PCR. Errors in the PCR process can lead to a high degree of randomization. The VH and / or VL libraries can be selected against target antigens or epitopes separately, in which case the binding to the individual domain is directly selected for each or as a whole.

A preferred method of preparing a specific double ligand comprises the use of a selection system in which a repertoire of variable domains is selected for binding to a first antigen or epitope and a repertoire of variable domains is selected for binding to a second antigen or epitope. The first and second selected variable domains are then combined and the specific double ligand is selected for binding to the first and second antigen or epitope. Closed forming ligands are selected for binding to

55 first or second antigen or epitope in isolation but not simultaneously.

A. Library vector systems

A variety of selection systems are known in the art that are suitable for use in the present invention. Examples of such systems are described below.

Lambda bacteriophage expression systems can be directly screened as bacteriophage plaques or as linogen colonies, as described above (Huse et al. (1989) Science, 246: 1275; Caton and Koprowski (1990) Proc. Natl. Acad. Sci. USA, 87; Mullinax et al. (1990) Proc. Natl. Acad. Sci. USA,

87: 8095; Persson et al. (1991) Proc. Natl Acad. Sci. U.S.A., 88: 2432) and can be used in the invention. While

5 These expression systems can be used for screening up to 106 different members of a library, they are not really suitable for screening large numbers (larger than 106 members).

In the construction of libraries, selection presentation systems are especially useful, allowing a nucleic acid to be linked to the polypeptide it expresses. As used herein, a selection presentation system is a system that allows the selection, by suitable means of presentation, of the members

10 individual library by binding to generic and / or target ligands.

Selection protocols for isolating desired members of large libraries, such as those typified by phage display techniques, are known in the art. Such systems, in which various peptide sequences are shown on the surface of filamentous bacteriophages (Scott and Smith (1990) Science, 249: 386), have been found useful for creating libraries of antibody fragments (and the nucleotide sequences that encode) for in vitro selection and amplification of specific antibody fragments that bind to target antigen (McCafferty et al., WO-92/01047). Nucleotide sequences encoding the VH and VL regions are linked to gene fragments encoding forward signals that direct them to the periplasmic space of E. coli and as a consequence the resulting antibody fragments are presented on the surface of the bacteriophage, usually as fusions with bacteriophage coating proteins (e.g., pIII or pVIII). Alternatively, the antibody fragments are presented externally in lambda phage capsids (phage bodies). An advantage of phage-based presentation systems is that, because they are biological systems, members of selected libraries can be amplified simply by growing the phage containing the selected library member in bacterial cells. In addition, since the nucleotide sequence encoding the polypeptide library member is contained in a phage or phagemid vector, the

Sequencing, expression and subsequent genetic manipulation are relatively simple.

Methods for the construction of bacteriophage antibody presentation libraries and lambda phage expression libraries are well known in the art (McCafferty et al. (1990) Nature, 348: 552; Kang et al. (1991) Proc. Natl. Acad. Sci. USA, 88: 4363; Clackson et al. (1991) Nature, 352: 624; Lowman et al. (1991) Biochemistry, 30: 10832; Burton et al. (1991) Proc. Natl. Acad. Sci USA, 88: 10134; Hoogenboom et al. (1991)

30 Nucleid Acid Res., 19: 4133; Chang et al. (1991) J. Immunol., 147: 3610; Breitling et al. (1991) Gene, 104: 147; Marks et al. (1991) above; Barbas et al. (1992) above; Hawkins and Winter (1992) J. Immunol., 22: 867; Marks et al., 1992, J. Biol. Chem., 267: 16007; Lerner et al. (1992) Science, 258: 1313, incorporated herein by reference).

A particularly advantageous approach has been the use of scFv phage libraries (Huston et al., 1988, Proc. Natl.

35 Acad. Sci U.S.A., 85: 5879-5883; Chaudhary et al. (1990) Proc. Natl Acad. Sci U.S.A., 87: 1066-1070; McCafferty et al. (1990) above; Clackson et al. (1991) Nature, 352: 624; Marks et al. (1991) J. Mol. Biol., 222: 581; Chiswell et al. (1992) Trends Biotech., 10: 80; Marks et al. (1992) J. Biol. Chem., 267). Several embodiments of scFv libraries presented in bacteriophage coating proteins have been described. Refinements of phage display strategies are also known, for example as described in WO96 / 06213 and

WO92 / 01047 (Medical Research Council et al.) And WO97 / 08320 (Morphosys), which are incorporated herein by reference.

Other systems for generating polypeptide libraries involve the use of cell-free enzyme machinery for in vitro synthesis of library members. In one method, RNA molecules are selected by alternating batches of selection against a target ligand and PCR amplification (Tuerk and Gold (1990) Science, 249: 505; 45 Ellington and Szostak (1990) Nature, 346: 818). A similar technique can be used to identify DNA sequences that bind to a predetermined human transcription factor (Thiesen and Bach (1990) Nucleic Acid Res., 18: 3203; Beaudry and Joyce (1992) Science, 257: 635; WO92 documents / 05258 and WO92 / 14843). Similarly, in vitro translation can be used to synthesize polypeptides as a method of generating large libraries. These methods, which generally comprise stabilized polysome complexes, are also described in

50 WO88 / 08453, WO90 / 05785, WO90 / 07003, WO91 / 02076, WO91 / 05058 and WO92 / 02536. Alternative phage-based presentation systems such as those described in WO95 / 22625 and WO95 / 11922 (Affymax) use the polysomes to present the polypeptides for selection.

An additional category of techniques involves the selection of repertoires in artificial compartments, which allow the union of a gene with its gene product. For example, a selection system is described in which the nucleic acids encoding desirable gene products can be selected in microcapsules formed by water-in-oil emulsions in WO99 / 02671, WO00 / 40712 and Tawfik & Griffiths (1998) Nature Biotechnol 16 (7), 652-6. The genetic elements that encode a gene product that has a desired activity are compartmentalized in microcapsules and then transcribed and / or translated to produce their respective gene products (RNA or protein) in the microcapsules. The genetic elements that produce a product

60 gene that has desired activity are subsequently classified. This strategy selects gene products of interest by detecting the desired activity by various means.

B. Library construction.

Libraries intended for selection may be constructed using techniques known in the art, for example as indicated above, or may be purchased from commercial sources. Libraries that are useful in the present invention are described, for example, in WO99 / 20749. Once a system of 5 vectors is chosen and one or more nucleic acid sequences that encode polypeptides of interest in the library vector are cloned, diversity can be generated within the cloned molecules by performing mutagenesis before expression; alternatively, the encoded proteins can be expressed and selected, as described above, before mutagenesis and additional selection runs are performed. The mutagenesis of nucleic acid sequences encoding structurally optimized polypeptides is performed by 10 standard molecular methods. Of particular use is the polymerase chain reaction, or PCR, (Mullis and Faloona (1987) Methods Enzymol., 155: 335, incorporated herein by reference). PCR, which employs multiple cycles of DNA replication catalyzed by a thermostable DNA-dependent DNA polymerase to amplify the target sequence of interest, is well known in the art. The construction of various antibody libraries has been addressed in Winter et al. (1994) Ann. Rev. Immunology 12, 433-55, and in

15 references cited in this document.

PCR is performed using template DNA (at least 1fg; more commonly, 1-1,000 ng) and at least 25 pmol of oligonucleotide primers; It may be advantageous to use a large amount of primer when the primer pool is very heterogeneous, since each sequence is represented by only a small fraction of the molecules in the pool, and the amounts become limiting in subsequent amplification cycles. A typical reaction mixture includes: 2 µl of DNA, 25 pmol of first oligonucleotide, 2.5 µl of 10X 1 PCR buffer (Perkin-Elmer, Foster City, CA), 0.4 ml of 1.25 dNTP M, 0.15 l (or 2.5 units) of Taq DNA polymerase (Perkin Elmer, Foster City, CA) and deionized water at a total volume of 25 l. Mineral oil is spread and PCR is performed using a programmable thermal cycler. The length and temperature of each stage of a PCR cycle, as well as the number of cycles, are adjusted according to the severity requirements in practice. The temperature of re-association and time are determined by the efficiency with which a primer is expected to be re-associated in a mold and the degree of absence of matching that will be tolerated; obviously, when nucleic acid molecules undergo amplification and mutagenesis simultaneously, no pairing is required, at least in the first round of synthesis. The ability to optimize the severity of the conditions of re-association of the primer is within the knowledge of one of ordinary skill in the art. A temperature of

30 re-association between 30 ° C and 72 ° C. The initial denaturation of the mold molecules normally takes place between 92 ° C and 99 ° C for 4 minutes, followed by 20-40 cycles consisting of denaturation (94-99 ° C for 15 seconds to 1 minute), reassociation ( temperature determined as set forth above; 1-2 minutes) and extension (72 ° C for 1-5 minutes, depending on the length of the amplified product). The final extension is generally 4 minutes at 72 ° C, and can be followed by an indefinite stage (0-24 hours) at 4 ° C.

35 C. Combination of individual variable domains

The domains useful in the invention, once selected, can be combined by various methods known in the art, including covalent and non-covalent methods.

Preferred methods include use of polypeptide linker groups, as described, for example, in relation to scFv molecules (Bird et al., (1988) Science 242: 423-426). An exhibition of the

40 suitable linker groups in Bird et al. Science 242, 423-426; Hudson et al., Journal Immunol. Methods 231 (1999) 177-189; Hudson et al., Proc Nat Acad Sci USA 85, 5879-5883. The linker groups are preferably flexible, which allows the interaction of the two individual domains. An example of a linker group is a linker group (Gly4Ser) n, in which n = 1 to 8, p. eg, 2, 3, 4, 5 or 7. Linker groups used in diabodies can also be used, which are less flexible (Holliger et al., (1993) PNAS (USA) 90: 6444-6448).

In one embodiment, the linker group employed is not a hinge region of immunoglobulins.

Variable domains can be combined using methods other than linker groups. For example, the use of disulfide bridges, provided through naturally occurring or designed cysteine residues, can be used to stabilize VH-VH, VL-VL or VH-VL dimers (Reiter et al., (1994) Protein Eng 7: 697-704) or by remodeling the interface between the variable domains to improve the "fit" and thus the stability of the

50 interaction (Ridgeway et al., (1996) Protein Eng. 7: 617-621; Zhu et al., (1997) Protein Science 6: 781-788).

Other techniques can be used to bind or stabilize variable domains of immunoglobulins, and in particular VH antibody domains, as appropriate.

The specific double ligands described herein may be in "closed" conformations in solution. A "closed" conformation is one in which the two domains (for example, VH and VL) are present

In associated form, such as an associated VH-VL pair that forms an antibody binding site. For example, scFv may be in a closed conformation, depending on the arrangement of the linker group used to link the VH and VL domains. If it is flexible enough to allow the association of the domains, or rigidly maintains them in the associated position, it is likely that the domains adopt a closed conformation.

Similarly, pairs of VH domains and pairs of VL domains may exist in a closed conformation. In general, this will depend on the close association of the domains, for example by a rigid linker group, in the ligand molecule. Ligands in a closed conformation may not bind to the molecule that increases the half-life of the ligand and to a second target molecule. Thus, the ligand will only normally join the second

5 target molecule in dissociation of the molecule that increases the half-life of the ligand.

On the other hand, the construction of VH / VH, VL / VL or VH / VL dimers without linker groups provides competition between domains.

On the other hand the ligands can be in an open conformation. In said conformation, the ligands may be simultaneously bound to the molecule that increases the half-life of the ligand and to the second target molecule.

10 Normally, the variable domains in an open configuration are (in the case of VH-VL pairs) sufficiently far from the domains that do not interact and form an antibody binding site and not to compete for binding to their respective epitopes. In the case of VH / VH or VL / VL dimers, domains are not forced to be joined by rigid linker groups. Naturally, such domain pairings will not compete for antigen binding or form an antibody binding site.

15 Fab fragments and whole antibodies will exist mainly in the closed conformation, although it will be observed that the existence of specific open and closed double ligands is likely in various equilibria under different circumstances. The binding of the ligand to a target will probably shift the equilibrium balance towards the open configuration. Thus, some ligands according to the invention may exist in two conformations in solution, one of which (the open form) can bind to two antigens or epitopes independently, while the

20 alternative conformation (the closed form) can only bind to an antigen or epitope; antigens or epitopes thus compete for ligand binding in this conformation.

Although the double specific ligand open form may thus exist in equilibrium with the closed form in solution, it is contemplated that equilibrium will favor the closed form; on the other hand, the open form can be sequestered by binding to the target in a closed conformation. Preferably, therefore, some specific double ligands

25 are present in a balance between two conformations (open and closed).

The specific double ligands described herein may be modified in order to favor an open or closed conformation. For example, stabilization of VH-VL interactions with disulfide bonds stabilizes the closed conformation. On the other hand, linker groups used to join domains, which include pairs of VH domain and VL domain, can be constructed so that the open form is favored;

For example, linker groups can sterically prevent the association of domains, for example by incorporation of large amino acid residues in appropriate places, or the design of a suitable rigid structure that will keep the domains physically separated.

D. Characterization of the specific double ligand.

The binding of the specific double ligand to its specific antigens or epitopes can be checked by methods that

35 will be familiar to those skilled in the art and that include ELISA. In a preferred embodiment, the binding is tested with monoclonal phage ELISA.

The phage ELISA technique can be performed according to any suitable procedure: an example protocol is explained below.

Phage populations produced in each selection run can be screened for binding by ELISA to

40 antigen or epitope selected, in order to identify "polyclonal" phage antibodies. Phages of the individual infected bacterial colonies of these populations can then be screened by ELISA to identify "monoclonal" phage antibodies. It is also convenient to screen soluble antibody fragments for antigen or epitope binding, which can also be done by ELISA using reagents, for example, against a C or N end tag (see for example Winter et al. ( 1994) Ann. Rev. Immunology

45 12, 433-55 and references cited therein).

The diversity of phage monoclonal antibodies can also be evaluated by gel electrophoresis of PCR products (Marks et al. 1991, above; Nissim et al. 1994 above), by probing (Tomlinson et al., 1992 J. Mol Biol. 227, 776) or by sequencing the vector DNA.

E. Structure of dobles double specific ligands ’.

As described above, an antibody is defined herein as an antibody (eg IgG, IgM, IgA, IgA, IgE) or fragment (Fab, Fv, Fv with disulfide bond, scFv, diabody) comprising the less a heavy and light chain variable domain, at least two heavy chain variable domains or at least two light chain variable domains. It can be obtained at least in part from any species that naturally produces an antibody, or created by recombinant DNA technology; either isolated from whey,

55 B lymphocytes, hybridomas, transfectomas, yeasts or bacteria).

In a preferred embodiment the double specific ligand comprises at least one individual heavy chain variable domain of an antibody and an individual light chain variable domain of an antibody, or two individual heavy or light chain variable domains. For example, the ligand may comprise a VH / VL pair, a pair of VH domains or a pair of VL domains.

5 The first and second variable domains of said ligand can be in the same polypeptide chain. Alternatively they can be in separate polypeptide chains. In the case that they are in the same polypeptide chain they can be linked by a linker group, which is preferably a peptide sequence, as described above.

The first and second variable domains can be associated covalently or noncovalently. If they are covalently associated, the covalent bonds can be disulfide bonds.

In case the variable domains are selected from repertoires of V genes selected for example using phage display technology as described herein, then these variable domains comprise a universal framework region, so that they can be recognized by a ligand generic specific as defined herein. The use of universal frames, generic ligands and the like is

15 described in WO99 / 20749.

When repertoires of V genes are used, the variation in the polypeptide sequence is preferably located in the structural bonds of the variable domains. The polypeptide sequences of each variable domain can be modified by shuffling DNA or by mutation in order to enhance the interaction of each variable domain with its complementary pair. DNA shuffling is known in the art and is taught, for example, in Stemmer, 1994,

20 Nature 370: 389-391 and US Pat. No. 6,297,053, which are incorporated herein by reference. Other mutagenesis methods are well known to those skilled in the art.

In a preferred embodiment the ’double specific ligand’ is a single stranded Fv fragment. In an alternative embodiment, the "specific double ligand" consists of a Fab format.

In a further aspect, the present disclosure provides a nucleic acid encoding at least one "specific double ligand" as defined herein.

One skilled in the art will observe that the two antigens or epitopes can bind simultaneously to the same antibody molecule. Alternatively, they can compete for binding to the same antibody molecule. For example, when the two epitopes are simultaneously bound, both variable domains of a specific double ligand are capable of independently binding to their target epitopes. When domains compete, the domain

The variable is capable of joining its target, but not at the same time that the other variable domain joins its similar target; or the first variable domain is able to bind to its target, but not at the same time as the second variable domain joins its similar target.

Variable domains can be obtained from antibodies directed against target antigens or epitopes. Alternatively, they can be obtained from a repertoire of individual antibody domains such as those expressed

35 on the surface of filamentous bacteriophages. The selection can be made as described below.

In general, nucleic acid molecules and artificial vector structures required for the realization of the present invention can be constructed and manipulated as set forth in standard laboratory manuals, such as Sambrook et al. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, USA

The manipulation of nucleic acids useful in the present invention is usually performed on recombinant vectors.

As used herein, "vector" refers to a discrete element that is used to introduce heterologous DNA into cells for expression and / or replication thereof. The methods by which these vectors are selected and constructed and subsequently used are well known to one skilled in the art. Numerous vectors are publicly available, including bacterial plasmids, bacteriophages, artificial chromosomes and 45 episomal vectors. Such vectors can be used for cloning and simple mutagenesis; alternatively a gene expression vector is used. A use vector according to the invention can be selected to accommodate a polypeptide coding sequence of a desired size, usually 0.25 kilobases (kb) at 40 kb or more in length. A suitable host cell is transformed with the vector after in vitro cloning manipulations. Each vector contains several functional components, which generally include a site of

50 cloning (or "linker polygroup"), an origin of replication and at least one selectable marker gene. If a given vector is an expression vector, it additionally has one or more of the following: enhancer element, promoter, signal sequences and transcription termination, each located in the vicinity of the cloning site, so that they are operatively linked to the gene encoding a ligand according to the invention.

Cloning and expression vectors generally contain nucleic acid sequences that allow the

Vector is replicated in one or more selected host cells. Normally in cloning vectors, this sequence is such that it allows the vector to replicate independently of the chromosomal DNA of the

host and includes replication sources or replicating sequences autonomously. Such sequences are well known for various bacteria, yeasts and viruses. The origin of replication of plasmid pBR322 is suitable for most gram-negative bacteria, the 2 micrometer plasmid origin is suitable for yeasts, and various viral sources (eg SV 40, adenovirus) are useful for cloning vectors in

5 mammalian cells In general, the origin of replication is not necessary for mammalian expression vectors unless they are used in mammalian cells capable of replicating high levels of DNA, such as COS cells.

Advantageously, a cloning or expression vector may contain a selection gene also referred to as a selectable marker. This gene encodes a protein necessary for the survival or growth of transformed host cells cultured in a selective culture medium. Host cells do not

10 transformed with the vector containing the selection gene will therefore not survive in the culture medium. Typical selection genes encode proteins that confer resistance to antibiotics and other toxins, e.g. ex. ampicillin, neomycin, methotrexate or tetracycline, complement auxotrophic deficiencies or provide critical nutrients not available in growth media.

Since the replication of vectors encoding a ligand according to the present invention is carried out in the most

15 and E. coli, a selectable marker of E. coli is used, for example, the β-lactamase gene that confers resistance to the antibiotic ampicillin. These elements can be obtained from E. coli plasmids, such as pBR322 or a pUC plasmid such as pUC18 or pUC19.

Expression vectors normally contain a promoter that is recognized by the host organism and that is operably linked to the coding sequence of interest. Said promoter can be inducible or

20 constitutive. The term "operably linked" refers to a juxtaposition where the components described are in a relationship that allows them to function as intended. A control sequence "operably linked" to a coding sequence is linked so that expression of the coding sequence is achieved under conditions compatible with the control sequences.

Promoters suitable for use with prokaryotic hosts include, for example, the systems

25 promoters of β-lactamase and lactose, alkaline phosphatase, the tryptophan (trp) promoter system and the hybrid promoters such as the tac promoter. Promoters for use in bacterial systems will also generally contain a Shine-Delgarno sequence operatively linked to the coding sequence.

Preferred vectors are expression vectors that allow the expression of a nucleotide sequence corresponding to a polypeptide library member. Thus, the selection with the first and / or 30 second antigen or epitope can be performed by separate propagation and expression of a single clone expressing the polypeptide library member or by the use of any selection presentation system. As described above, the preferred selection presentation system is the presentation of bacteriophages. Thus, phage or phagemid vectors can be used, e.g. ex. pIT1 or pIT2. The forward sequences useful in the invention include pelB, stII, ompA, phoA, bla and pelA. An example is phagemid vectors that have an E. 35 coli replication origin (for double-stranded replication) and also a phage replication origin (for single-stranded DNA production). The manipulation and expression of such vectors is well known in the art (Hoogenboom and Winter (1992) above; Nissim et al. (1994) above). Briefly, the vector contains a β-lactamase gene to confer selectivity in the phagemid and a 5 'lac promoter of an expression cassette consisting of (in terminal N to C) a pelB forward sequence (which directs the expressed polypeptide to the periplasmic space), a multiple cloning site (for cloning the nucleotide version of the library member), optionally, one or more peptide tags (for detection), optionally, one or more TAG stop codons and the pIII phage protein. Thus, using several suppressor and non-suppressor strains of E. coli and with the addition of glucose, iso-propylthio-β-D-galactoside (IPTG) or an auxiliary phage, such as VCS M13, the vector is able to replicate as a plasmid without expression, of producing large quantities of the polypeptide library member alone or of

Phage, some of which contain at least one copy of the pIII polypeptide fusion on its surface.

The construction of vectors encoding ligands according to the invention employs conventional binding techniques. Isolated vectors or DNA fragments are cleaved, adapted and reattached in the desired manner to generate the required vector. If desired, the analysis to confirm that the correct sequences are present in the constructed vector can be performed in a known manner. Suitable methods for constructing expression vectors, preparing in vitro transcripts, introducing DNA into host cells and performing analyzes to evaluate expression and function are known to those skilled in the art. The presence of a gene sequence in a sample is detected, or its amplification and / or expression is quantified by conventional methods, such as Southern or Northern analysis, Western blotting, point hybridization of DNA, RNA or proteins, in situ hybridization, immunocytochemistry or analysis. of nucleic acid sequences or protein molecules. The experts in the

Techniques will easily understand how to modify these methods, if they wish.

Structure of multispecific ligands of closed conformation

According to one aspect of the second configuration of the present description, the two or more non-complementary epitope binding domains are linked so that they are in a closed conformation as defined herein. Advantageously, they can also be attached to a skeleton that can, such as

Alternatively, or in addition to a linker group described herein, facilitate the formation and / or maintenance of the closed conformation of epitope binding sites with each other.

(I) Skeletons

Skeletons can be based on immunoglobulin molecules or can have a non-immunoglobulin origin

5 as indicated above. Preferred immunoglobulin skeletons as defined herein include any one or more of those selected from the following: an immunoglobulin molecule comprising at least (i) the CL domain (subclass kappa or lambda) of an antibody; or (ii) the CH1 domain of an antibody heavy chain; an immunoglobulin molecule comprising the CH1 and CH2 domains of an antibody heavy chain; an immunoglobulin molecule comprising the

10 domains CH1, CH2 and CH3 of a heavy chain of antibodies; or any of the subset (ii) in conjunction with the CL domain (subclass kappa or lambda) of an antibody. A hinge region domain can also be included. Such combinations of domains can emulate, for example, natural antibodies, such as IgG or IgM,

or fragments thereof, such as Fv, scFv, Fab or F (ab ’) 2 molecules. Those skilled in the art will understand that this list is not intended to be exhaustive.

15 (II) Protein frameworks

Each epitope binding domain comprises a protein shell and one or more CDRs that are involved in the interaction of the specific domain with one or more epitopes. Advantageously, an epitope binding domain according to the present invention comprises three CDRs. Suitable protein frameworks include any of those selected from the group consisting of the following: those based on immunoglobulin domains,

20 based on fibronectin, those based on affibody molecules, those based on CTLA4, those based on chaperones such as GroEL, those based on lipocalin and those based on bacterial Fc receptors SpA and SpD. Those skilled in the art will note that this list is not intended to be exhaustive.

F:

Frameworks for use in the construction of specific double ligands

i.

 Selection of the main chain conformation

25 Members of the immunoglobulin superfamily share all similar folds for their polypeptide chains. For example, although antibodies are very diverse in terms of their primary sequence, the comparison of crystallographic sequences and structures has revealed that, contrary to expectations, five of the six antigen-binding bonds of antibodies (H1, H2, L1 , L2, L3) adopt a limited number of main chain conformations, or canonical structures (Chothia and Lesk (1987) J. Mol. Biol., 196: 901; Chothia et al. (1989)

30 Nature, 342: 877). The analysis of the loop lengths and key residues has therefore allowed to predict the main chain conformations of H1, H2, L1, L2 and L3 present in most human antibodies (Chothia et al. (1992) J. Mol Biol., 227: 799; Tomlinson et al. (1995) EMBO J., 14: 4628; Williams et al. (1996) J. Mol. Biol., 264: 220). Although the H3 region is much more diverse in terms of sequence, length and structure (due to the use of D segments), it also forms a limited number of main chain conformations for lengths

35 short loops that depend on the length and presence of particular residues, or types of residues, at key positions in the loop and the main structure of the antibody (Martin et al. (1996) J. Mol. Biol., 263: 800; Shirai et al. (1996) FEBS Letters, 399: 1).

The specific double ligands described herein are advantageously assembled from domain libraries, such as VH domain libraries and / or VL domain libraries. On the other hand, the specific double ligands 40 can themselves be provided in the form of libraries. In one aspect, libraries of specific double ligands and / or domains in which certain loop lengths and key residues have been chosen are designed to ensure that the main chain conformation of the members is known. Advantageously, they are real conformations of immunoglobulin superfamily molecules present in nature, to minimize the chances of them being nonfunctional, as discussed above.

45 The V segments of germ lines act as a suitable basic framework for building libraries of antibodies or T-cell receptors; other sequences are also used. Variations can occur at a very low frequency, so that a small number of functional members may have an altered main chain confirmation, which does not affect their function.

The theory of canonical structures is also used to evaluate the number of different conformations of

50 main chain encoded by ligands, to predict the conformation of main chain based on ligand sequences and to choose residues for diversification that do not affect the canonical structure. It is known that, in the human Vκ domain, the L1 loop can adopt one of four canonical structures, the L2 loop has a single canonical structure and 90% of the human Vκ domains adopt one of four or five canonical structures for the L3 loop (Tomlinson et al. (1995) above); thus, in the Vκ domain alone, different can be combined

55 canonical structures to create a range of different main chain conformations. Since the Vλ domain encodes a different range of canonical structures for the L1, L2 and L3 loops and that the Vκ and Vλ domains can be paired with any VH domain that can encode several canonical structures for the H1 and H2 loops, the number of combinations of canonical structures observed for these five ties is very broad. This

It implies that the generation of diversity in the conformation of the main chain may be essential for the production of a wide range of binding specificities. However, when building an antibody library based on a single known main chain conformation it has been found, contrary to expectations, that diversity in the main chain conformation is not necessary to generate enough

5 diversity to target substantially all antigens. Even more surprisingly, the only main chain conformation does not have to be a consensus structure; a single natural occurrence conformation can be used as the basis for the entire library. Thus, in a preferred aspect, the double specific ligands possess a single known main chain conformation.

The only main chain conformation that is chosen is preferably common among molecules of the type of

10 immunoglobulin superfamily in question. A conformation is common when it is observed that it is adopted by a significant number of naturally occurring molecules. Accordingly, in a preferred aspect, the natural occurrence of the different main chain conformations for each binding loop of an immunoglobulin domain is considered separately, and then a naturally occurring variable domain having the combination of conformations of Main chain desired for different ties. If there is not

15 none available, the closest equivalent can be chosen. It is preferable that the combination of desired main chain conformations for the different loops is created by selecting germline gene segments encoding the desired main chain conformations. It is more preferable that the selected germline gene segments are frequently expressed in nature, and it is most preferable that they be expressed with the maximum frequency of all natural germline gene segments.

In the design of specific double ligands or libraries thereof, the incidence of the different main chain conformations for each of the six antigen binding bonds can be considered separately. For H1, H2, L1, L2 and L3, a given conformation is chosen that is adopted by between 20% and 100% of the antigen-binding bonds of naturally occurring molecules. Normally, its observed incidence is greater than 35% (ie between 35% and 100%) and, ideally, greater than 50% or even greater than 65%. Dice

Since the vast majority of H3 loops do not have canonical structures, it is preferable to select a main chain conformation that is common among those loops that present canonical structures. For each of the ties, the conformation that is observed most frequently in the natural repertoire is therefore selected. In human antibodies, the most popular canonical structures (CS) for each loop are the following: H1 -CS 1 (79% of the expressed repertoire), H2 -CS 3 (46%), L1 -CS 2 of Vκ (39% ), L2 -CS 1 (100%), L3 -CS 1 of Vκ (36

30%) (the calculation assumes a κ: λ ratio of 70:30, Hood et al. (1967) Cold Spring Harbor Symp. Quant. Biol., 48: 133). For H3 loops that have canonical structures, the most common seems to be the CDR3 length (Kabat et al. (1991) Sequences of proteins of immunological interest, US Department of Health and Human Services) of seven wastes with a salt bridge from residue 94 to residue 101. There are at least 16 human antibody sequences in the EMBL data library with the required H3 length and key residues to form this

35 conformation and at least two crystallographic structures in the protein database that can be used as a basis for antibody modeling (2cgr and 1tet). The most frequently expressed germline gene segments of this combination of canonical structures are segment VH 3-23 (DP-47), segment JH JH4b, segment Vκ O2 / O12 (DPK9) and segment Jκ Jκ1. VH segments DP45 and DP38 are also suitable. These segments can therefore be used in combination as a basis to build a library with

40 the desired individual main chain conformation.

Alternatively, instead of choosing the individual main chain conformation based on the natural occurrence of the different main chain conformations for each of the joint loops in isolation, the natural occurrence of main chain conformation combinations is used as the basis to choose the individual main chain conformation. In the case of antibodies, for example, the natural occurrence of combinations of canonical structures for two, three, four, five or for the six antigen binding bonds can be determined. In this case, it is preferable that the conformation chosen is common in naturally occurring antibodies and it is most preferable that it is observed with the highest frequency in the natural repertoire. Thus, in human antibodies, for example, when natural combinations of the five antigen binding bonds, H1, H2, L1, L2 and L3 are considered, the most frequent combination of canonical structures is determined and then

50 is combined with the most popular conformation for the H3 loop, as the basis for choosing the individual main chain conformation.

ii. Diversification of the canonical sequence

Once selected several known main chain conformations or, preferably a single known main chain conformation, the specific double ligands or libraries for use in the invention

55 can be constructed by modifying the binding site of the molecule in order to generate a repertoire with structural and / or functional diversity. This means that variants are generated so that they have enough diversity in their structure and / or function so that they are capable of providing a range of activities.

The desired diversity is normally generated by varying the selected molecule in one or more positions. The positions that will be changed can be chosen at random or preferably selected. The variation can then be achieved by randomization, during which the resident amino acid is substituted by any natural or synthetic amino acid or analog thereof, which produces a very large number of variants or

replacing the resident amino acid with one or more of a defined subset of amino acids, to produce a more limited number of variants.

Several methods to introduce such diversity have been reported. Error-prone PCR can be used (Hawkins et al. (1992) J. Mol. Biol., 226: 889), chemical mutagenesis (Deng et al. (1994) J. Biol. Chem., 269: 9533) or strains 5 Bacterial mutators (Low et al. (1996) J. Mol. Biol., 260: 359) to introduce random mutations in the genes encoding the molecule. Methods for mutating selected positions are also known in the art and include the use of absence of paired oligonucleotides or degenerated oligonucleotides, with or without the use of PCR. For example, several libraries of synthetic antibodies have been created by directing mutations to the antigen binding bonds. The H3 region of a human tetanus toxoid binding FAB has been randomized to create a range of new binding specificities (Barbas et al. (1992) Proc. Natl. Acad. Sci. USA, 89: 4457). Random or semi-random H3 and L3 regions have been added to gene segments V of germ lines to produce large libraries with non-mutated framework regions (Hoogenboom & Winter (1992) J. Mol. Biol., 227: 381; Barbas et al. (1992) Proc. Natl. Acad. Sci. USA, 89: 4457; Nissim et al. (1994) EMBO J., 13: 692; Griffiths et al. (1994) EMBO J., 13: 3245; De Kruif et al. (1995) J. Mol. Biol., 248: 97). This diversification has been extended to

15 include part or all of the other antigen binding bonds (Crameri et al. (1996) Nature Med., 2: 100; Riechmann et al. (1995) Bio / Technology 13: 475; Morphosys, WO97 / 08320 , higher).

Since randomization of the loops has the potential to create approximately 1015 structures for H3 alone and a similarly large number of variants for the other five loops, it is not feasible to use current transformation technology or even cell-free systems to produce a library that represents all

20 possible combinations. For example, in one of the largest libraries built to date, 6 x 1010 different antibodies were generated, which is only a fraction of the potential diversity for a library of this design (Griffiths et al. (1994) above).

In a preferred embodiment, the only residues that are directly involved in the creation or modification of the desired function of the molecule are diversified. For many molecules, the function will bind to the target and by

Both diversity should be concentrated at the target binding site, while avoiding changing the residues that are crucial to the overall packaging of the molecule or to maintain the chosen main chain conformation.

Diversification of the canonical sequence in its application to antibody domains

In the case of specific double antibody ligands, the binding site for the target is most often

30 antigen binding site. Thus, in a highly preferred aspect, the description provides libraries of or for the assembly of specific double ligands of antibodies in which the only residues at the antigen binding site vary. These residues are extremely diverse in the repertoire of human antibodies and are known to form contacts in high resolution antibody / antigen complexes. For example, in L2 it is known that positions 50 and 53 are diverse in naturally occurring antibodies and it is observed that they form contact with the

35 antigen. Instead, the conventional approach would be to diversify all residues in the corresponding complementarity determination region (CDR1) as defined by Kabat et al. (1991, above), about seven wastes compared to the two diversified in the library for use as described herein. This represents a significant improvement in terms of the functional diversity required to create a range of antigen binding specificities.

40 In nature, antibody diversity is the result of two processes: somatic recombination of germline V, D and J gene segments to create a new primary repertoire (the so-called germline and junction diversity) and somatic hypermutation of the resulting rearranged V genes. Analysis of human antibody sequences has shown that diversity in the primary repertoire focuses on the center of the antigen binding site while somatic hypermutation extends diversity to regions in the periphery.

45 of the antigen binding site that are highly conserved in the primary repertoire (see Tomlinson et al. (1996) J. Mol. Biol., 256: 813). This complementarity has probably evolved as an effective strategy for the search of the sequence space and, although apparently unique for the antibodies, can easily be applied to other polypeptide repertoires. The residues that are varied constitute a subset of those that form the binding site for the target. Different subsets (including overlap) of waste on site

50 target binding are diversified in different phases during selection, if desired.

In the case of an antibody repertoire, an initial ‘new’ repertoire is created in which some, but not all, of the residues at the antigen binding site are diversified. As used herein in this context, the term "new" refers to antibody molecules that have no predetermined target. These molecules resemble those encoded by the immunoglobulin genes of a person who has not undergone

55 to immune diversification, as is the case of fetuses and newborns, whose immune systems have not yet been faced with a wide variety of antigenic stimuli. This repertoire is then selected against a range of antigens or epitopes. If needed, greater diversity can be introduced outside the diversified region in the initial repertoire. This mature repertoire can be selected for a modified function, specificity or affinity.

The description herein provides two repertoires of different new binding domains for the construction of specific double ligands, or a new library of specific double ligands, in which part

or all of the residues at the antigen binding site are varied. The "primary" library mimics the natural primary repertoire, with diversity limited to residues at the center of the antigen-binding site that are diverse in 5 gene segments V of the germ line (germ line diversity) or are diversified during recombination process (binding diversity). Residues that are diversified include, but are preferably not limited to, H50, H52, H52a, H53, H55, H56, H58, H95, H96, H97, H98, L50, L53, L91, L92, L93, L94 and L96. In the "somatic" library, diversity is limited to residues that diversify during the recombination process (binding diversity) or are such that somatically they are very mutated). Residues that diversify include, but are preferably not limited to: H31, H33, H35, H95, H96, H97, H98, L30, L31, L32, L34 and L96. For all residues listed as suitable for diversification in these libraries it is known that they form contact with one or more antibody-antigen complexes. Since in the two libraries not all residues at the antigen binding site are varied, additional diversity is incorporated during selection by modifying the remaining residues, if desired. It will be obvious to a person skilled in the art that he can

Any subset of any of these residues (or additional residues comprising the antigen binding site) is used for initial diversification and / or subsequent diversification of the antigen binding site.

In the construction of libraries for use described herein, diversification of chosen positions is usually achieved at the level of nucleic acids, altering the coding sequence that specifies the sequence of the polypeptide so that a series of possible amino acids can be incorporated. (all 20

or a subset of them) in that position. Using the nomenclature of the IUPAC, the most versatile codon is NNK, which encodes all amino acids as well as the stop codon MARK. The NNK codon is preferably used to introduce the required diversity. Other codons that achieve the same ends are also used, including the NNN codon, which leads to the production of the additional stop codons TGA and TAA.

A characteristic of the side chain diversity at the human antibody antigen binding site is a

25 pronounced bias and favorable to certain amino acid residues. If the amino acid composition of the ten most diverse positions in each of the VH, Vκ and Vλ regions is added, more than 76% of the diversity of side chains comes from only seven different residues, which are serine (24%) , tyrosine (14%), asparagine (11%), glycine (9%), alanine (7%), aspartate (6%) and threonine (6%). This bias towards hydrophilic residues and small residues that can provide flexibility of the main chain probably reflects the evolution of surfaces that are predisposed to binding to a wide range of antigens or epitopes and can help explain the required promiscuity of antibodies in the primary repertoire .

As it is preferable to mimic this distribution of amino acids, the distribution of amino acids in the positions to be varied preferably mimics that observed at the antibody antigen binding site. This bias in amino acid substitution that allows the selection of certain polypeptides (not just antibody polypeptides)

35 against a range of target antigens is easily applied to any repertoire of polypeptides. There are several methods to skew the distribution of amino acids in the position to be varied (which includes the use of trinucleotide mutagenesis, see WO97 / 08320), of which the preferred method, due to the ease of synthesis, is the use of conventional degenerate codons. Comparing the amino acid profile encoded by all combinations of degenerate codons (with single, double, triple and quadruple degeneration in equal proportions in each position) with the use of natural amino acids it is possible to calculate the most representative codon. The codons (AGT) (AGC) T, (AGT) (AGC) C and (AGT) (AGC) (CT). that is, DVT, DVC and DVY, using the IUPAC nomenclature, are the closest to the desired amino acid profile: they code for 22% serine and 11% tyrosine, asparagine, glycine, alanine, aspartate, threonine and cysteine. Preferably, therefore, libraries are constructed using the codon DVT, DVC or DVY in each of the diversified positions.

45 G: Antigens capable of increasing the half-life of ligands

The specific double ligands described herein, in a configuration thereof, are capable of binding to one or more molecules that can increase the half-life of the ligand in vivo. Normally, these molecules are polypeptides that appear naturally in vivo and that resist degradation or elimination by endogenous mechanisms that eliminate unwanted material from the body. For example, the molecule that increases the half-life of the organism can be selected from the following:

Extracellular matrix proteins; for example collagen, laminins, integrins and fibronectin. Collagens are the main proteins of the extracellular matrix. At present, approximately 15 types of collagen molecules are known, present in different parts of the body, e.g. ex. Type I collagen (which accounts for 90% of the body's collagen) present in bone, skin, tendon, ligaments, cornea, internal organs or type collagen

55 II present in the cartilage, the intervertebral disc, the notocordium, the vitreous humor of the eye.

Proteins present in the blood, including: plasma proteins such as fibrin, macroglobulin α-2, serum albumin, fibrinogen A, fibrinogen B, serum amyloid protein A, heptaglobin, profilin, ubicuitin, uteroglobulin and β2-microglobulin;

Enzymes and inhibitors such as plasminogen, lysozyme, cystatin C, alpha-1-antitrypsin and trypsin inhibitor

pancreatic Plasminogen is the inactive precursor of the trypsin-like serine protease plasmin. It is normally in circulation through the bloodstream. When plasminogen is activated and converted to plasmin, it displays a potent enzyme domain that dissolves fibrinogen fibers that entangle blood cells in a clot. This phenomenon is called fibrinolysis.

5 Immune system proteins, such as IgE, IgG, IgM.

Transport proteins such as retinol binding protein, microglobulin α-1.

Defensins such as beta-defensin 1, neutrophil defensins 1, 2 and 3.

Proteins present in the blood brain barrier or in nervous tissues, such as the melanocortin receptor, myelin, ascorbate transporter.

10 Pharmaceutical agent-transferrin specific ligand fusion protein (see US5977307); brain capillary endothelial cell receptor, transferrin, transferrin receptor, insulin, insulin-like growth factor receptor 1 (IGF 1), insulin-like growth factor receptor 2 (IGF 2), insulin receptor .

Proteins located in the kidney, such as polycystin, type IV collagen, K1 organic anion transporter, Heymann antigen.

Proteins located in the liver, for example alcohol dehydrogenase, G250.

Blood coagulation factor X

Α1 antitrypsin

HNF 1α

20 Proteins located in the lung, as a secretory component (binds to IgA).

Proteins located in the heart, for example HSP 27. It is associated with dilated cardiomyopathy.

Proteins located in the skin, for example keratin.

Bone-specific proteins, such as bone morphogenic proteins (BMP), which are a subset of the transforming growth factor β superfamily that shows osteogenic activity. Examples include

BMP-2, -4, -5, -6, -7 (also referred to as osteogenic protein (OP-1) and -8 (OP-2).

Tumor-specific proteins, including human trophoblast antigen, herceptin receptor, estrogen receptor, cathepsins p. ex. Cathepsin B (present in the liver and spleen).

Disease-specific proteins, such as antigens expressed only in activated T lymphocytes: including LAG-3 (lymphocyte activation gene), osteoprotegerin ligand (OPGL) see Nature 402, 304-309; 1999, 30 OX40 (a member of the TNF receptor family, expressed in activated T lymphocytes and the only costimulatory T lymphocyte molecule known to be specifically regulated by an increase in the cells producing type I human T lymphocyte leukemia virus (HTLV-I)). See J Immunol. 2000 Jul 1; 165 (1): 263-70; Metalloproteases (associated with arthritis / cancers), which include Drosophila CG6512, human paraplechin, human FtsH, human AFG3L2, murine ftsH; Angiogenic growth factors, which includes growth factor of

35 acid fibroblasts (FGF-1), basic fibroblast growth factor (FGF-2), vascular endothelial growth factor / vascular permeability factor (VEGF / VPF), transforming growth factor-a (TGF a), factor of tumor necrosis-alpha (TNF-α), angiogenin, interleukin-3 (IL-3), interleukin-8 (IL-8), platelet-derived endothelial growth factor (PD-ECGF), placental growth factor (P1GF ), growth factor derived from midline platelets-BB (PDGF), fractalcin.

40 HSP of stress proteins (heat shock proteins) normally present in the intracellular environment. When they are present extracellularly, it is an indicator that the cell has died and has poured its contents. This unscheduled cell death (necrosis) only occurs when as a result of trauma, disease or injury and therefore in vivo, extracellular HSPs trigger an immune system response that will fight infection and disease. A specific double ligand that is bound to extracellular HSP can

45 be located at a disease site.

Proteins involved in the transport of Fc

Brambell receiver (also known as FcRB)

This Fc receptor has two functions, both potentially useful for childbirth.

The functions are

(one)

The transport of IgG from the mother to the child through the placenta

(2)

the protection of IgG from degradation and thus prolonging the serum half-life of IgG. It is believed that the receptor recycles IgG from the endosome.

5 See Holliger et al., Nat Biotechnol 1997 Jul; 15 (7): 632-6.

The ligands described herein can be designed to be specific for the above targets without requiring any direct or half-life increase in vivo. For example, the ligands according to the invention may be specific for targets selected from those described above that are tissue specific, thereby allowing specific targeting of the tissue of the specific double ligand, or

10 dAb monomer that binds to a therapeutically relevant tissue-specific target, regardless of possible increases in half-life, although these may occur. On the other hand, when the dAb ligand or monomer is directed to the kidney or liver, it can redirect the dAb ligand or monomer to an alternative in vivo clearance pathway (for example, the ligand can be directed from the clearance in the liver to the clearance in the kidney).

As described above, the ligands described herein that comprise a domain

Individual variable as defined herein may be selected to be specific to a target and preferably may have the added attribute of increasing the half-life of a target in vivo, although not necessary. A specific double ligand can be composed of an individual variable domain of heavy chain of antibodies that has a specificity of binding to a first epitope or antigen, and also of a single variable domain of light chain of antibodies that has a specificity of binding to a second

The epitope or antigen, in which one or both of the antigens can be serum albumin, or one or both of the epitopes is an epitope or epitopes of serum albumin. In one embodiment, the two serum albumin epitopes are the same, in another embodiment, each serum albumin epitope is different.

In addition to these double specific ligands that have the attribute of increasing the half-life of a target in vivo, other structural forms of ligands that have or consist of at least one individual variable domain as defined herein are described herein. which has the attribute of increasing the half-life of a target binding ligand in vivo, e.g. eg, by binding to serum albumin. For example, the ligand may consist of, or contain, an individual variable domain of monomers as defined herein that binds to serum albumin; or the ligand may be in a form comprising multiple individual variable domains as defined herein, in which one or more of the individual variable domains binds to serum albumin, that is, a multimer. Both the multimer and the monomer may comprise other entities in addition to the one or more individual variable domains that bind to serum albumin, e.g. eg, in the form of a fusion protein and / or a conjugate. Said fusion protein is preferably a single chain of polypeptides and may comprise, for example, two or more linked individual variable domains as defined herein; the linked individual variable domains may be identical or may be different from each other. 35 Such entities include p. eg, one or more additional individual variable domains as defined herein, which have a specificity for an antigen or epitope other than serum albumin, and / or one or more drugs, and / or one or more binding domains target that have a specificity to an antigen or epitope other than serum albumin and that are not individual variable domains as defined herein. Said multimer may have multiple valences with respect to its domain or individual variable domains, e.g. eg

40 univalent, divalent, trivalent, tetravalent. Said multimer may be in the form of an IgG structure or a specific double ligand as defined herein, as well as other structures such as IgM, IgE, IgD or IgA, and / or fragments thereof, which include but are not limited to to fragments such as scFv, Fab, Fab 'fragments, etc. The ligand can be modified to contain additional moieties, such as a fusion protein or a conjugate.

The present invention provides a ligand comprising an individual variable domain of antibodies, in

Wherein said individual variable antibody domain comprises the amino acid sequence dAb7h11 having the amino acid sequence depicted in Figure 24, Sheet 1, or an amino acid sequence that is identical thereto at least 95%, wherein the Beta T half-life of the individual variable domain in cynomolgus is at least 50% of the beta T half-life of cynomolgus serum albumin in cynomolgus.

A ligand having one or more individual variable chain domains is described herein.

Light antibody and wherein the one or more individual variable domains of the light chain of antibodies specifically bind to serum albumin and have an amino acid sequence of an individual variable domain of the light chain of antibodies selected from, but preferably not limited to , the group: dAb2, dab4, dAb7, dAb11, dAb12, dAb13, dAb15, dAb16, dAb17, dAb18, dAb19, dAb21, dAb22, dAb23, dAb24, dAb25, dAb26, dAb27, dAb30, dAb31, dAb33, dAb34, DAB35, dAb38, dAb41, dAb46, dAb47, dAb52, dAb53, dAb54, dAb55, dAb56,

55 drdAb7m12, dAb7m16, dAb7m26, dAb7r1, dAb7r3, dAb7r4, dAb7r5, dAb7r7, dAb7r8, dAb7r13, dAb7r14, dAb7r15, dAb7r16, dAb7r17, dAb7r18, dAb7r19, dAb7h1, dAb7h2, dAb7h6, dAb7h7, dAb7h8, dAb7h9, dAb7h10, dAb7h11, dAb7h12 , dAb7h13, dAb7h14, dAb7p1, dAb7p2, and a sequence that is at least identical to it in 80%, or up to, inclusive, 85%, 90%, 95%, 96%, 97 %, 98% or 99% identity to it.

The ligand embodiments described above and herein also include those that have a structure comprising an IgG structure that has any combination of one, or two of the above specific double ligands, and / or individual variable domains comprising frameworks of no immunoglobulins Said immunoglobulin structure may have several combinations of individual variable domains of antibodies, including an IgG structure containing four individual variable domains of heavy chain antibodies, or an IgG structure containing four individual variable domains of light chain antibodies, thus as an IgG structure containing two pairs of chains, each pair containing an individual variable domain of heavy chain of antibodies and a single variable domain of light chain of antibodies. In addition to these IgG structures, the ligands described herein may contain one

or more monomers of an individual variable domain, which includes but is preferably not limited to the individual variable domains listed above, in which if the ligand contains more than one of these individual variable domains, the individual variable domains may be identical to each other or not identical to each other.

Embodiments of ligands comprising one or more individual variable domains include, but are preferably not limited to, the dAb described herein, double specific monomers comprising at least one individual variable domain, double specific IgG molecules containing monomers single chain antibodies and multivalent IgG molecules comprising single chain antibody monomers as described herein. These embodiments may further comprise a binding site for a generic ligand. The generic ligand may include, but is preferably not limited to, protein A, protein L and protein G. In said double specific ligand, which includes the specific double ligands present in an IgG format, one or the individual variable domains specifically bind to an epitope

or antigen with a dissociation constant (Kd) that can be selected from, but preferably not limited to, 1 x 10-3 Momenos, 1 x 10-4 Momenos, 1 x 10-5 Momenos, 1 x 10- 6 Moments, 1 x 10-7 Momenos, 1 x 10-8 M or less and 1 x 10-9 M or less, as determined, for example, by resonance by surface plasmons. Said double specific ligand, which includes those specific double ligands present in an IgG format, may

25 also contain one or more entities that include, but preferably are not limited to a label, a brand and a drug. Said double specific ligand, which includes the specific double ligands present in an IgG format, as well as a multimeric ligand containing one or more monomers of the individual variable domains listed above, may be present in a kit, and in a composition, which includes a pharmaceutical composition, which contains the specific double ligand and a vehicle thereof.

Similarly, for a ligand comprising one or more individual variable domains as described herein, which includes a monomeric ligand and a multimeric ligand as defined above, the one or more individual variable domains specifically bind. to an epitope or antigen with a dissociation constant (Kd) that may be selected from, but preferably not limited to, 1 x 10-3 M or less, 1 x 10-4 M or less, 1 x 10-5 M or less, 1 x 10-6 M or less, 1 x 10-7 M or less, 1 x 10-8 M or less and 1 x

10-9 M or less, as determined, for example, by resonance by surface plasmons. Said ligand may also contain one or more entities that include, but preferably are not limited to a label, a brand and a drug. Said ligand may be present in a kit, a composition, which includes a pharmaceutical composition, which contains the ligand and a vehicle thereof.

H: Use of multispecific ligands according to the second configuration of the description

The multispecific ligands according to the method of the second configuration of the present description can be used in therapeutic and prophylactic applications in vivo, in vitro and in vivo diagnostic applications, in vitro assay and reagent applications and the like. For example, antibody molecules can be used in antibody-based assay techniques, such as ELISA techniques, according to methods known to those skilled in the art.

As mentioned above, multispecific ligands are used in diagnostic, prophylactic and therapeutic procedures. Multispecific antibodies are used for diagnostic purposes in Western analysis and detection of proteins in situ by standard immunohistochemical procedures; For use in these applications, ligands can be labeled according to techniques known in the art. In addition, said antibody polypeptides can be used preparatively in affinity chromatography procedures, when they complex with a chromatographic support, such as a resin. All these techniques are well known to one skilled in the art.

Diagnostic uses of multispecific closed conformation ligands include homogeneous assays for analytes that take advantage of the ability of multispecific closed conformation ligands to bind to two competing targets, so that two targets cannot be joined simultaneously (a closed conformation),

or alternatively its ability to join two targets simultaneously (an open conformation).

55 The manufacturers of diagnostic and research test systems used in drug discovery and treatment have looked forward to a true homogeneous immunoassay format. The main diagnostic markets include human tests in hospitals, consultations and medical clinics, commercial reference laboratories, blood banks and homes, non-human diagnoses (for example food tests, water tests, environmental tests, biodefense and tests). veterinarians), and finally research (which includes drug development; basic research and academic research).

Currently all these markets use immunoassay systems that are built around chemiluminescence, ELISA, fluorescence or in rare cases radioimmunoassay. Each of these test formats requires a separation step (separation of bound and unbound reagents). In some cases, several stages of separation are required. The inclusion of these additional steps adds reagents and automation, takes time and influences the final test result. In human diagnosis, the separation stage can be automated, which masks the problem, but does not eliminate it. Robotics, additional reagents, additional incubation times, and the like add considerable cost and complexity. In the development of drugs, such as high-performance screening, in which literally millions of samples are tested at a time, with very low levels of the test molecule, the inclusion of additional separation stages can eliminate the ability to perform a screening . However, if separation is avoided, too much noise is created in the reading. Thus, there is a need for a true homogeneous format that provides sensitivities in the range that can be obtained in the present assay formats. Advantageously, an assay has fully quantitative readings with high sensitivity and a large dynamic range. Sensitivity is an important requirement, as well as reducing the amount of sample required. These two characteristics are features offered

15 a homogeneous system. It is very important in the tests at the point of attention, and in the development of drugs in which the samples are precious. Heterogeneous systems, such as those currently available in the art, require large amounts of sample and expensive reagents.

Applications for homogeneous assays include cancer tests, in which the largest trial is that of prostate specific antigen, used in screening in prostate cancer men. Other applications include fertility tests, which provides a series of tests for women trying to get pregnant in which beta-hcg is included for pregnancy. Tests for infectious diseases, which include hepatitis, HIV, rubella and other viruses and microorganisms and sexually transmitted diseases. Blood bank tests are used, especially tests for HIV, hepatitis A, B, C, not A no B. Therapeutic drug follow-up tests include the levels of follow-up of drugs prescribed in patients in terms of efficacy and

25 to prevent toxicity, for example digoxin for arrhythmias and phenobarbital levels in psychotic cases; Theophylline for asthma. On the other hand, diagnostic tests are useful in drug tests, such as tests for cocaine, marijuana and the like. Metabolic tests are used to measure thyroid function, anemia and other psychological functions and disorders.

On the other hand, the homogeneous immunoassay format is useful in the manufacture of standard clinical chemistry assays. The inclusion of immunoassays and chemistry assays in the same instrument is highly advantageous in diagnostic tests. Suitable chemical tests include tests for glucose, cholesterol, potassium and the like.

An additional important application for homogeneous immunoassays is the discovery and development of drugs: high-performance screening includes tests of combinatorial chemistry libraries against targets in

35 ultra high volume. The signal is detected, and then the positive groups are divided into smaller groups, and finally tested in cells and then in animals. Homogeneous tests can be used in all these types of tests. In drug development, especially in animal studies and clinical trials, intense use of immunoassays is made. Homogeneous trials greatly accelerate and simplify these procedures. Other applications include food and beverage tests: meat and other food tests for E. coli, salmonella, etc .; water tests, which include tests on aquatic plants of all types of pollutants including E. coli; and veterinary tests.

In an extensive embodiment, a binding assay is described herein comprising a detectable agent that is bound to a closed-forming multi-specific ligand according to the invention, and whose detectable properties are altered by the binding of an analyte to said multi-specific ligand. of conformation

45 closed. Said assay can be configured in several different ways, each of which takes advantage of the above properties of the multispecific closed conformation ligands.

The assay is based on the direct or indirect displacement of an agent by the analyte, which results in a change in the detectable properties of the agent. For example, in which the agent is an enzyme that is capable of catalyzing a reaction that has a detectable titration criteria, said enzyme may be bound by the ligand so as to obstruct its active site, thereby inactivating the enzyme. The analyte, which is also bound by the closed-forming multispecific ligand, displaces the enzyme, making it active through the release of the active site. The enzyme is then able to react with a substrate, to provide the increase to a detectable event. In an alternative embodiment, the ligand can bind to the enzyme outside the active site, which influences the conformation of the enzyme and thus modifies its activity. For example, the structure of the active site can

55 may be limited by ligand binding, or the binding of cofactors necessary for activity can be avoided.

The physical implementation of the assay can take any form known in the art. For example, the closed-forming multi-specific ligand / enzyme complex can be provided in a test strip; The substrate can be provided in a different region of the test strip, and a solvent containing the analyte is allowed to migrate through the ligand / enzyme complex, displacing the enzyme, and carrying it to the region of the substrate to produce a signal. Alternatively, the ligand / enzyme complex can be provided on a test rod.

or in another solid phase, and immersed in an analyte / substrate solution, releasing the enzyme in the solution in

Response to the presence of analyte.

Since each analyte molecule potentially releases an enzyme molecule, the assay is quantitative, with the strength of the signal generated at a given time depending on the concentration of analyte in the solution.

Additional configurations that use the analyte in a closed conformation are possible. For example, the ligand

Multi-specific closed conformation can be configured to bind an enzyme in an allosteric site, thereby activating the enzyme. In said embodiment, the enzyme is active in the absence of analyte. The addition of the analyte displaces the enzyme and eliminates allosteric activation, thus inactivating the enzyme.

In the context of the above embodiments that employ enzymatic activity as a measure of analyte concentration, the activation or inactivation of the enzyme refers to an increase or decrease in the activity of the enzyme, measured as the capacity of the enzyme. of catalyzing a signal generation reaction. For example, the enzyme can catalyze the conversion of an undetectable substrate into a detectable form thereof. For example, horseradish peroxidase is widely used in the art together with chromogenic or chemiluminescent substrates, which are commercially available. The level of increase or decrease in enzyme activity can be between 10% and 100%, for example 20%, 30%, 40%, 50%, 60%, 70%, 80

15% or 90%; in the case of an increase in activity, the increase may be more than 100%, that is, 200%, 300%, 500% or more, or it may not be measurable as a percentage if the base activity of the inhibited enzyme is undetectable.

In a further configuration, the multispecific closed conformation ligand can bind to the substrate of an enzyme / substrate pair, rather than the enzyme. The substrate is therefore not available for the enzyme until it is

20 releases the closed-forming multispecific ligand through the junction of the analyte. The implementations for this configuration are as for the configurations that bind to the enzyme.

On the other hand, the assay can be configured to bind to a fluorescent molecule, such as a fluorescein or other fluorophore, in a conformation so that the fluorescence is inactivated at ligand binding. In this case, the binding of the analyte to the ligand will displace the fluorescent molecule, thus producing a signal. The alternatives to

Fluorescent molecules that are useful in the present invention include luminescent agents, such as luciferin / luciferase, and chromogenic agents, which include agents commonly used in immunoassays such as HRP.

The therapeutic and prophylactic uses of multispecific ligands prepared according to the description involve the administration of ligands according to the invention to a recipient mammal, such as a human being. Multi-specificity

30 may allow antibodies to bind to the multimeric antigen with great avidity. Multispecific ligands can allow cross-linking of two antigens, for example in the recruitment of cytotoxic T lymphocytes to mediate the destruction of tumor cell lines.

Substantially pure ligands or binding proteins thereof, for example dAb monomers, of at least 90 to 95% homogeneity are preferred for administration to a mammal, and most preferred is 98 to 99% or more homogeneity. for pharmaceutical uses, especially when the mammal is a human being. Once purified, partially or even homogeneously as desired, the ligands can be used for diagnostic purposes.

or therapeutic (including extracorporeal form) or in the development and performance of test procedures, immunofluorescent stains and the like (Lefkovite and Pernis, (1979 and 1981) Immunological Methods, Volumes I and II, Academic Press, NY).

40 Ligands will normally find use in the prevention, suppression or treatment of inflammatory conditions, allergic hypersensitivity, cancer, bacterial or viral infection and autoimmune disorders (including, but not limited to, type I diabetes, asthma, multiple sclerosis, arthritis rheumatoid, systemic lupus erythematosus, Crohn's disease and myasthenia gravis).

In the present application, the term "prevention" implies the administration of the protective composition before the

45 induction of the disease. "Suppression" refers to the administration of the composition after an inducing event, but before the clinical onset of the disease. "Treatment" implies the administration of the protective composition after the symptoms of the disease have manifested.

Animal model systems are available that can be used to screen the efficacy of the antibodies or their binding proteins in the protection or treatment of the disease. Methods for 50 systemic lupus erythematosus (SLE) tests in sensitive mice are known in the art (Knight et al. (1978) J. Exp. Med., 147: 1653; Reinersten et al. (1978) New Eng. J. Med., 299: 515). Myasthenia gravis (MG) is tested in female SJL / J mice by inducing disease with soluble AchR protein of another species (Lindstrom et al. (1988) Adv. Immunol., 42: 233). Arthritis is induced in a sensitive strain of mice by injection of type II collagen (Stuart et al. (1984) Ann. Rev. Immunol., 42: 233). A model has been described by which adjuvant arthritis is induced in sensitive rats by injection of mycobacterial heat shock proteins (Van Eden et al. (1988) Nature, 331: 171). Thyroiditis is induced in mice by the administration of thyroglobulin as described (Maron et al. (1980) J. Exp. Med., 152: 1115). Insulin-dependent diabetes mellitus (DMID) occurs naturally or can be induced in certain strains of mice such as those described by Kanasawa et al. (1984)

Diabetologia, 27: 113. The EAE in mouse and rat acts as a model for MS in humans. In this model, demyelinating disease is induced by administration of myelin basic protein (see Paterson (1986) Textbook of Immunopathology, Mischer et al., Eds., Grune and Stratton, New York, pp. 179-213; McFarlin et al . (1973) Science, 179: 478: and Satoh et al. (1987) J. Immunol., 138: 179).

5 A ligand comprising an individual variable domain, or a composition thereof, that specifically binds to vWF, e.g. eg, human vWF, a vWF A1 domain, the activated vWF A1 domain or the vWF A3 domain, may further comprise a thrombolytic agent. This thrombolytic agent may be covalently or non-covalently linked to an individual variable domain, in particular to an individual variable domain of antibodies, by covalent or non-covalent means known to one skilled in the art. The media does not

Covalent agents include the interaction of proteins such as biotin / streptavidin, or through an immunoconjugate. Alternatively, the thrombolytic agent may be administered simultaneously, separately or in sequence with respect to a ligand consisting of or comprising an individual variable domain that binds to vWF or a vWF domain as described above, or a composition thereof. The thrombolytic agents according to the invention may include, for example, staphylokinase, tissue plasminogen activator, streptokinase,

15 single chain streptokinase, urokinase and acylo plasminogen streptokinase complex.

Also described herein are invasive medical devices coated by an individual variable domain, or a ligand comprising an individual variable domain, or a composition thereof, or an individual variable domain that comes from a screening method described herein. . Non-limiting examples of the devices include surgical intubation, occlusion devices, prosthetic devices. The

20 application for such devices includes surgical procedures that require modulation of platelet-mediated aggregation around the site of invasion (eg a device coated with an individual variable domain that specifically binds to vWF) or a modulation of inflammation (eg a device coated with an individual variable domain that specifically binds TNF alpha).

In general, the present ligands will be used in purified form together with pharmacologically carriers.

25 appropriate. Typically, these vehicles include aqueous or alcoholic / aqueous solutions, emulsions or suspensions, whichever includes saline medium and / or with buffered medium. Parenteral vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride and Ringer's solution with lactate. Suitable physiologically acceptable adjuvants, if it is necessary to retain a complex of suspended polypeptides, can be chosen from thickeners such as carboxymethyl cellulose, polyvinylpyrrolidone, gelatin and alginates.

30 Intravenous vehicles include fluid and nutrient replacement and electrolyte replacement agents, such as those based on Ringer's dextrose. Preservatives and other additives may also be present, such as antimicrobials, antioxidants, chelating agents and inert gases (Mack (1982) Remington’s Pharmaceutical Sciences, 16th edition).

The ligands of the present invention can be used as compositions administered separately or in

35 conjunction with other agents. These may include various immunotherapeutic drugs, such as cyclosporine, methotrexate, adriamycin or cisplatin, and immunotoxins. Pharmaceutical compositions may include "cocktails" of various cytotoxic or other agents in conjunction with the ligands of the present invention, or even combinations of ligands according to the present invention that have different specificities, such as selected ligands using different target antigens or epitopes, whether included. or not in reserve before administration.

The route of administration of the pharmaceutical compositions according to the invention may be any of those commonly known to those skilled in the art. For therapy, which includes without limitation immunotherapy, the selected ligands thereof of the invention can be administered to any patient according to standard techniques. Administration may be by any appropriate way, including parenteral, intravenous, intramuscular, intraperitoneal, transdermal, pulmonary, or, appropriately, by

45 direct infusion with a catheter. The dose and frequency of administration will depend on the age, sex and condition of the patient, the concurrent administration of other drugs, contraindications and other parameters that will be taken into account by the clinician.

The ligands of the present invention can be lyophilized for storage and reconstituted in a suitable vehicle before use. This technique has proven effective in conventional immunoglobulins and

50 lyophilization and reconstitution techniques known in the art can be employed. Those skilled in the art will note that lyophilization and reconstitution may lead to varying degrees of loss of antibody activity (eg with conventional immunoglobulins, IgM antibodies usually have greater loss of activity than IgG antibodies) and may have You adjust the levels of use upwards to compensate.

The compositions containing the present ligands or a cocktail thereof can be administered for prophylactic and / or therapeutic treatments. In some therapeutic applications, an amount adequate to achieve at least partial inhibition, suppression, modulation, destruction or some other measurable parameter, of a population of selected cells is defined as a "therapeutically effective dose". The amounts needed to achieve this dosage will depend on the severity of the disease and the condition.

general of the patient's own immune system, but in general they will be between 0.005 and 5.0 mg of ligand, e.g. ex. antibody, receptor (eg a T-cell receptor) or binding protein thereof per kilogram of body weight, with doses of 0.05 to 2.0 mg / kg / dose being more commonly used. For prophylactic applications, compositions containing the present ligands or cocktails thereof can also be administered in

5 similar or slightly lower dosages.

The treatment performed using the compositions described herein is considered "effective" if one or more symptoms are reduced (eg, by at least 10% or at least one point on the clinical assessment scale), with respect to to said symptoms present before treatment, or in relation to said symptoms in an individual (human being or animal model) not treated with said composition. The symptoms will obviously vary according to the disease or target disorder, but can be measured by a technician or an expert clinician. Such symptoms can be measured, for example, by monitoring the level of one or more biochemical indicators of the disease or disorder (e.g., values of an enzyme or metabolite correlated with the disease, numbers of affected cells, etc.)., by monitoring clinical manifestations (eg, inflammation, tumor size, etc.)., or by an accepted clinical assessment scale, for example, the Extended Disability Status Scale (for 15 multiple sclerosis), the Irvine Questionnaire of Inflammatory Bowel Disease (32-point assessment that assesses the quality of life with respect to bowel function, systemic symptoms, social function and emotional state, with a score between 32 and 224, so that assessments higher indicate better quality of life), the Rheumatoid Arthritis Quality of Life Scale or the accepted clinical assessment scale as known in the field. A sustained reduction (e.g., one day or more, preferably more time) in the

20 symptoms of the disease or disorder at least 10% or at one or more points on a given clinical scale is indicative of "effective" treatment. Similarly, prophylaxis performed using a composition as described herein is "effective" if the onset or severity of one or more symptoms is delayed, reduced or eliminated with respect to such symptoms in a similar individual (human being). or animal model) not treated with the composition.

A composition containing a ligand or a cocktail thereof according to the present invention can be used in prophylaxis or therapy scenarios to contribute to the alteration, inactivation, destruction or elimination of a selected population of target cells in a mammal. In addition, the repertoires of selected polypeptides described herein can be used extracorporeally or selectively in vitro to effectively destroy, deplete or otherwise eliminate a population of target cells from a heterogeneous collection of

30 cells The blood of a mammal can be combined extracorporeally with the ligands, e.g. ex. antibodies, cell surface receptors or binding proteins thereof in which unwanted cells are destroyed or removed by other means of blood for return to the mammal according to standard techniques.

Selection and characterization of ligands comprising an individual variable domain for serum albumin binding from a set of species

35 A ligand can comprise one or more individual variable domains, e.g. eg, individual variable domains of immunoglobulins and / or individual variable domains of non-immunoglobulins, in which at least one of the individual variable domains is specifically bound to serum albumin of a human being, as well as non-human species. In one embodiment, the individual variable domain specifically binds to only serum albumin that is endogenous to humans. In another embodiment, the individual variable domain specifically binds to

40 serum albumin of a non-human species. Alternatively, the individual variable domain specifically binds to a serum albumin that is endogenous to humans, as well as to a serum albumin that is endogenous to one or more non-human species. As a non-limiting example, said individual variable domain can specifically bind to human endogenous serum albumin and cynomolgus, or human and rat endogenous serum albumin, or human and mouse serum albumin, or human and pig serum albumin.

Alternatively, the individual variable domain is specifically linked to two or more serum albumins of two or more non-human species. As used herein, serum albumin may be expressed by an endogenous gene of a species, that is, natural serum albumin, and / or by a recombinant equivalent thereof. In one embodiment, the serum albumin includes fragments, analogs and derivatives of natural and recombinant serum albumin. Such serum albumin fragments include fragments containing domain I, domain

II and / or domain III, or combinations of one or two or more of each of domains I, II and III of serum albumin, preferably human serum albumin. Domain II of serum albumin is preferred as a target for the individual variable domain as defined herein. Other preferred combinations are Domain I and Domain II; Domain I and Domain III; Domain II and Domain III; and Domain I alone; Domain II alone; and Domain III alone; and Domain I and Domain II and Domain III. In one embodiment, the serum albumin is serum albumin.

A recombinant expressed exogenously in a non-human host, such as an animal host, or a unicellular host such as yeast or bacteria.

The species of which serum albumin is endogenous includes any species that expresses endogenous serum albumin, which includes, but is not preferably limited to, the human, mouse, murine, rat, cynomolgus, porcine, host species of dog, cat, horse, goat and hamster. In some cases it

60 excludes endogenous serum camel or llama albumin.

In one embodiment, this individual variable domain is specifically bound to serum albumin that is endogenous.

from a first non-human species with a Kd value that is comprised between 10 times of the Kd value with which it specifically binds (i.e., with cross reaction) to at least one serum albumin that is endogenous to a second non-human species. Alternatively, this individual variable domain is specifically bound to serum albumin that is endogenous to the first non-human species with a Kd value that is 15, 20, 25, 30,

50 or up to about 100 times the Kd value with which it specifically binds (i.e., cross-reactive) to at least one serum albumin that is endogenous to the second non-human species. In some embodiments, the Kd may be between 300 nM and about 5 pM. In other embodiments, the individual variable domain is specifically bound to serum albumin with a Koff of at least 5 x 10-1, S-1, 5 x 10-2 S-1, 5 x 10-3 S-1, 5 x 10-4 S-1, 5 x 10-5 S-1, 5 x 10-6 S-1, 5 x 10-7 S-1, 5 x 10-8 S-1, 5 x 10-9 S- 1.5 x 10-10 S-1, or less, preferably with a Koff between 1 x 10-6 S-1 and 1 x 10-8 S-1.

For example, said ligand can include an individual variable domain of immunoglobulins, in which the individual variable domain of immunoglobulins is specifically bound to human serum albumin and mouse serum albumin, and in which the beta T half-life of the ligand is substantially the same than the beta T half-life of mouse serum albumin in a mouse host. In one version of said ligand, the epitope binding domain contains a non-immunoglobulin framework that is specifically bound to human serum albumin and mouse serum albumin, and wherein the beta T half-life of the ligand is substantially the same as the half-life. Mouse serum albumin beta T in a mouse host. The phrase "substantially the same" means that the ligand has a beta T half-life in a mouse host that is at least 50% of that of the mouse serum albumin in a mouse host, which is at least 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97%, 98 %, 99%, 100%, 101%, 102%, 105%, 110%, 125%, and up to 150% than the beta T half-life of mouse serum albumin in a host mouse The non-immunoglobulin framework may optionally include fragments of an individual variable domain of antibodies, such as one or more of the CDR regions of a variable antibody domain, which includes an individual variable domain of antibodies that has a beta T half-life in a human host. which is at least 50%, 55

25%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97% , 98%, 99%, 101%, 102%, 105%, 110%, 125%, or up to 150% than that of the human serum albumin beta T half-life in a human host.

For example, one embodiment is an individual variable domain, in which the individual variable domain is specifically bound to human serum albumin and rat serum albumin, and in which the beta T half-life of the ligand is substantially the same as the beta T half-life. of rat serum albumin in a rat host. In one version of said ligand, the individual variable binding domain contains a non-immunoglobulin framework that is specifically bound to human serum albumin and rat serum albumin, and wherein the beta T half-life of the ligand is substantially the same as the beta half-life. Rat serum albumin T in a rat host. The phrase "substantially the same" means that the ligand has a beta T half-life in a mouse host.

35 which is at least 50% than that of rat serum albumin in a rat host, which is up to 55%, 60%, 65%, 70%, 75%, 80%, 85 %, 90%, 92%, 94%, 95%, 96%, 97%, 98%, 99%, 100%, 101%, 102%, 105%, 110%, 125%, up to 150% than the beta T half-life of rat serum albumin in a rat host. The non-immunoglobulin framework may optionally include fragments of an individual variable domain of antibodies, such as one or more of the CDR regions of a variable antibody domain.

For example, a ligand may include an individual variable domain of immunoglobulins, in which the individual variable domain of immunoglobulins is specifically bound to human serum albumin and porcine serum albumin, and in which the beta T half-life of the ligand is substantially the same as the half-life beta T of swine serum albumin in a pig host. In a version of a ligand, the epitope binding domain contains a

45 non-immunoglobulin framework that is specifically bound to human serum albumin and porcine serum albumin, and wherein the beta T half-life of the ligand is substantially the same as the porcine serum albumin beta T half-life in a porcine host. The phrase "substantially the same" means that the ligand has a beta T half-life in a porcine host that is at least 50% than that of porcine serum albumin in a porcine host, which is up to 55%, 60%, the 65%, 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97%, 98%, 99% , 100%, 101%, 102%, 105%, 110%, 125%, up to 150% than that of the swine serum albumin beta T half-life in a porcine host. The non-immunoglobulin framework may optionally include fragments of an individual variable domain of antibodies, such as one or more of the CDR regions of a variable antibody domain, which includes an individual variable domain of antibodies.

For example, a ligand can include an individual variable domain of immunoglobulins, in which the individual variable domain of immunoglobulins is specifically bound to human serum albumin and cynomolgus serum albumin, and in which the beta T half-life of the ligand is substantially the same as the beta T half-life of cynomolgus serum albumin in a cynomolgus host. In one version of a ligand, the serum albumin binding domain contains a non-immunoglobulin framework that is specifically bound to human serum albumin and cynomolgus serum albumin, and wherein the beta T half-life of the ligand is substantially the same as the Cynomolgus serum albumin beta T half-life in a cynomolgus host. The phrase "substantially the same" means that the ligand has a beta T half-life in a cynomolgus host that is at least 50% of that of cynomolgus serum albumin in a cynomolgus host, which is up to 55%, 60% , 65%, the

70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97%, 98%, 99%, 100% , 101%, 102%, 105%, 110%, 125%, or up to 150% than that of the beta T half-life of cynomolgus serum albumin in a cynomolgus host.

The non-immunoglobulin framework may optionally include fragments of an individual variable domain of 5 antibodies, such as one or more of the CDR regions of a variable antibody domain.

In one embodiment, a ligand and / or double specific ligand contains an individual variable domain that is specifically bound to serum albumin that is endogenous to humans, has a beta T half-life in a human host that is at least 50% of that of human serum albumin in a human host, up to 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97%, 98%, 99 10%, 100%, 101%, 102%, 105%, 110%, 125% or up to 150% of the half-life beta Tde human serum albumin in a human host. In a preferred embodiment, the individual variable domain that is specifically bound to serum albumin that is endogenous to a non-human species, has a beta T half-life in its respective non-human host that is at least 50% than that of serum albumin. human in its respective non-human host, up to 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 15 92%, 94%, 95%, 96%, 97%, 98%, 99%, 100%, 101%, 102%, 105%, 110%, 125%, or up to 150% that of the beta T half-life of non-human serum albumin in its respective non-human host. In a preferred embodiment, the individual variable domain that is specifically bound to serum albumin that is endogenous to the human being, and that also specifically binds to serum albumin of at least one non-human species, has a beta T half-life in a human host that it is up to 55%, 60%, 65%, 20 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96% , 97%, 98%, 99%, 101%, 102%, 105%, 110%, 125%, or up to 150% human serum albumin in a human host, and a half-life beta T in the non-human host that is up to 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 92%, 94%, 95%, 96%, 97%, 98%, 99%, 100%, 101%, 102%, 105%, 110%, 125%, or up to 150% of serum non-human albumin in its respective non-human host. In some embodiments, the beta T half-life of the individual variable domain that is specifically bound to serum albumin may be only 2 hours to and that includes 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours. , 10 hours, 11 hours, 12 hours, 14 hours, 16 hours, 18 hours, 20 hours, 22 hours, 1 day, 2 days, 3 days, 4 days, 4 days, 6 days, 8 days, 10 days, 12 days, 14 days, 16 days, 18 days, up to 21 days or more. In a human host, as well as a non-human host such as a pig, cynomolgus, rat, murine, mouse host, the 30 beta T half-life of the individual variable domain that is specifically bound to serum albumin may be only 2 hours to and which includes 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 14 hours, 16 hours, 18 hours, 20 hours, 22 hours, 1 day, 2 days, 3 days, 4 days, 4 days, 6 days, 8 days, 10 days, 12 days, 14 days, 16 days, 18 days, up to 21 days, or more. Other preferred beta T half-lives of a ligand comprising an individual variable domain that is specifically bound to serum albumin include: in a monkey host of about 3 to about 5, 6, 7 or 8 days, which includes only 2 hours, up to and including 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 14 hours, 16 hours, 18 hours, 20 hours, 22 hours, 1 day, 2 days, 3 days, 4 days, 4 days, 6 days, 8 days, 10 days, 12 days, 14 days, 16 days, 18 days, up to 21 days. In a rat or mouse host, the beta T half-life of the individual variable domain that is specifically bound to serum albumin may be only 40 hours to about 75 hours, and includes

40 only 2 hours up to and including 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 14 hours, 16 hours, 18 hours, 20 hours, 22 hours, 1 day, 2 days, 3 days, 4 days, 4 days, 6 days, 8 days, 10 days, 12 days, 14 days, 16 days, 18 days, up to 21 days.

An individual variable domain mentioned above may also additionally specifically bind human serum albumin with a Koff of at least 5 x 10-1, S-1, 5 x 10-2 S-1, 5 x 10-3 S-1 , 5 x 10-4 S-1, 5 x 10-5 S-1, 5 x 45 10-6 S-1, 5 x 10-7 S-1, 5 x 10-8 S-1, 5 x 10 -9 S-1, 5 x 10-10 S-1, or less, preferably with a Koff between

S-1 S-1

1 x 10-6 and 1 x 10-8. Individual variable domains that specifically bind human serum albumin and serum albumin that is endogenous to a non-human species can also bind to serum albumin that is endogenous to a third, fourth, fifth, sixth, seventh, eighth, ninth or tenth non-human species In a non-limiting embodiment, the individual variable domain that is specifically bound to serum albumin

50 human and rat serum albumin, is also specifically bound to cynomolgus serum albumin. In another non-limiting embodiment, the individual variable domain that is specifically bound to human serum albumin and mouse serum albumin, is also specifically linked to cynomolgus serum albumin.

As described herein, a ligand containing an individual variable domain (monomer) or more than one individual variable domain (multimer, fusion protein, conjugate and double specific ligand as defined herein) that is attached specifically to serum albumin, it may further comprise one or more entities selected from, but preferably not limited to, a label, a brand, an additional individual variable domain, a dAb, an antibody, and an antibody fragment, a label and a drug. One or more of these entities may be located at the COOH end or at the N end or at the N end and at the COOH end of the ligand comprising the individual variable domain, (either an individual variable domain of 60 immunoglobulins or not immunoglobulins). One or more of these entities may be located at the COOH end, or the N end, or at the N end and the COOH end of the individual variable domain that specifically binds to serum albumin of the ligand containing an individual variable domain (monomer) or more than one individual variable domain (multimer, fusion protein, conjugate and double specific ligand as defined in the

present memory). Non-limiting examples of brands that can be placed at one or both of these ends include an HA, his or a myc mark. Entities, which include one or more brands, labels and drugs, may be linked to the ligand containing an individual variable domain (monomer) or more than one individual variable domain (multimer, fusion protein, conjugate and specific double ligand as defined in the

5), which binds to serum albumin, either directly or through linker groups as described in a subsequent separate section.

A ligand that contains an individual variable domain (monomer) or more than one individual variable domain (multimer, fusion protein, conjugate and double specific ligand as defined herein) that is specifically bound to serum albumin, or that binds specifically to human serum albumin and at least one non-human serum albumin, it can specifically bind to one or more of Domain I and / or Domain II and / or Domain III of human serum albumin, as described in detail below. In addition to containing one or more individual variable domains (e.g., an individual variable domain of immunoglobulin-binding serum immunoglobulins or a single non-variable domain of serum albumin-binding immunoglobulins) that is specifically bound to a serum albumin, such as human serum albumin, or that specifically binds to human serum albumin and at least one non-human serum albumin, the ligand may contain one or more additional domains capable of specifically binding an antigen and / or epitope other than serum albumin, the antigen being selected or epitope among the group consisting of any animal protein, which includes cytokines, and / or antigens derived from microorganisms, pathogens, unicellular organisms, insects, viruses, algae and plants. These one or more additional domains that bind to a residue other than serum albumin may be a

20 non-immunoglobulin binding domain, an individual non-immunoglobulin variable domain and / or an individual immunoglobulin variable domain.

An isolated nucleic acid encoding any of the ligands described herein is also encompassed herein, e.g. eg, a ligand containing an individual variable domain (monomer) or more than one individual variable domain (eg, multimer, fusion protein, conjugate and double specific ligand as defined herein) that is attached specifically to serum albumin, or that specifically binds to human serum albumin and at least one non-human serum albumin, or functionally active fragments thereof. Also understood herein is a vector and / or an expression vector thereof, a host cell comprising the vector, e.g. eg, a plant or animal cell and / or a cell line transformed with a vector, a method of expression and / or production of one or more ligands containing an individual variable domain (monomer) or more than an individual variable domain (e.g., multimer, fusion protein, conjugate and double specific ligand as defined herein) that is specifically bound to serum albumin,

or fragment or fragments thereof encoded by said vectors, which in some cases include the culture of the host cell so that the one or more ligands or fragments thereof are expressed and optionally recover the ligand containing an individual variable domain ( monomer) or more than one individual variable domain (e.g., multimer, fusion protein, conjugate and double specific ligand as defined herein) that is specifically bound to serum albumin, of the cell culture medium host Contact methods of a ligand described herein with serum albumin, including serum albumin and / or non-human serum albumin or albumin, and / or one or more other targets of serum albumin, in which the targets include are also included. biologically active molecules, and include animal proteins, cytokines as indicated above, and include methods in which contacting is in vitro as well as the administration of any of the ligands described herein to an individual host animal or cell in vivo and / or ex vivo. Preferably, the administration of ligands described herein comprising an individual variable domain (immunoglobulin or non-immunoglobulin) directed to serum albumin and / or albumin or non-human serum albumin, and one or more domains directed to one or more different targets. of serum albumin, the half-life, which includes the half-life beta T, of the antidian ligand will increase. Nucleic acid molecules encoding the individual domain containing ligands or fragments thereof, including functional fragments thereof, are described herein. Vectors herein encoding nucleic acid molecules are described, which include but are preferably not limited to expression vectors, such as host cells of a germ line or organism containing

50 one or more of these expression vectors. Methods for producing any individual domain containing ligands are also described, which includes, but is preferably not limited to any of the nucleic acids, vectors and host cells mentioned above.

Serum albumin epitope mapping

Serum albumins of mammalian species have a similar structure, which contains three domains

55 predominant with a folding and pattern of similar disulfide bonds, as highlighted in Figure 25. It is believed that proteins should come from two tandem duplication episodes, and subsequent diversification of residues.

The structure of human serum albumin has been resolved by X-ray crystallography, with / without a variety of bound ligands:

60 • Atomic structure and chemistry of human serum albumin. I have XM, Carter DC.

•

Nature 1992; 358: 209-15. Errata in: Nature 1993; 364: 362.

•

Atomic structure and chemistry of human serum albumin. I have XM, Carter DC; J Mol Biol. 2001; 314: 955-60.

•

Crystal structures of human serum albumin complexed with monounsaturated and polyunsaturated fatty acids. Petitpas I, Grune T, Bhattacharya AA, Curry S .; J Biol Chem. 2001; 276: 22804-9.

5 Human serum albumin has been shown to be a heart-shaped molecule. The individual domains, called I, II and III, are predominantly helical, and each is composed of two subdomains, called IA, IB, IIA, 2B, IIIA and IIIB. They are joined by flexible random propellers.

A ligand containing one or more individual variable domains that are specifically linked to a human serum albumin II domain is described herein. The individual variable domain can be a single variable domain of VH antibodies. The individual variable domain can be an individual variable domain of VHH antibodies. The individual variable domain VH can be an individual variable domain VH3. The individual variable domain VH3 can be an individual variable domain VH3 human. The ligand can alternatively,

or additionally, include an individual variable domain that is an individual variable domain of VKappa antibodies, which includes one of the following: DOM7h-1, DOM7h-8, DOM7h-9, DOM7h-11, DOM7h-12, DOM7h-13, DOM7h -14.

15 DOM7r-3 and DOM7r-16, or an individual variable domain of VKappa antibodies having a domain that has an amino acid sequence of approximately 80%, 85%, 90%, 91%, 92%, the 93%, 94%, 95%, 96%, 97%, 98%, 99% identity or higher.

The individual variable domain of antibodies may include a set of four framework regions (FR) of Kabat, which are encoded by the VH antibody, preferably gene segments of framework germline antibodies

20 VH3. The VH3 frame is selected from the group consisting of DP47, DP38 and DP45. The individual variable domain of antibodies may include a set of four Kabat framework regions (FR) that are encoded by an antibody VL framework, preferably a gene segment of VKappa framework germline antibodies. Preferably, the Kappa frame is DPK9.

The ligand that contains one or more individual variable domains that specifically bind to Domain II of

Human serum albumin may further include one or more domains capable of specifically binding to a distinct serum albumin moiety, and may further comprise one or more entities that include one or more of a label, a brand and a drug. The one or more domains capable of specifically binding to a different serum albumin moiety may be an individual variable domain of immunoglobulins. Also described herein is a ligand that contains one or more individual variable domains that are specifically bound to a

30 domain II of human serum albumin, including the domain a framework of non-immunoglobulins and CDR1, CDR2 and / or CDR3 regions, or in which at least one of the CDR1, CDR2 and / or CDR3 regions is of an individual variable domain of an individual variable domain of antibodies that binds a domain II of human serum albumin. Non-immunoglobulin frameworks include, but are preferably not limited to, CTLA-4, lipocalin, staphylococcal protein A (SPA), Affibody ™, Avimers ™, GroEL and fibronectin.

The ligand containing one or more individual variable domains that are specifically bound to a human serum albumin II domain includes those domains that specifically bind human serum albumin with a Kd less than or equal to 300 nM. The ligand containing one or more individual variable domains that are specifically bound to a human serum albumin domain II may further comprise one or more entities that include one or more of a label, a brand and a drug. The brand may include one or more of

40 HC or myc marks at the C end or HA or myc marks at the N end.

Serum albumin has a long serum half-life compared to other serum proteins, along with a positive relationship between serum concentration and fractional catabolic rates (i.e., the higher the concentration of SA, the more the amount is degraded), a property that it shares with IgG Recently it has been discovered that IgG and serum albumin share a recycling mechanism, mediated by the neonatal FcRn Fc receptor. FcRn is a member of the MHC family of type I, composed of a heterodimer of the membrane-anchored FCRGT chain, and a non-membrane-bound beta-2 microglobulin. Inactivated mutant mice of FcRn or beta-2 microglobulin do not express functional FcRn, and show an increased biosynthesis rate of serum albumin (~ 20% increase), and a higher serum albumin catabolism, which leads to a lower serum level of the 40% serum albumin, with a shorter half-life (Chaudhry et al 2005). In humans, it has been proven

50 that mutations in beta-2 microglobulin provide very low functional FcRn levels and ultimately IgG deficiency and hypoalbuminemia, characterized by a reduced serum half-life of HSA (Wani et al. 2006, PNAS).

Without wishing to be bound by any theory, the proposed mechanism for salvage mediated by FcRn is as follows:

55 1. Plasma proteins are subjected to pinocytosis by cells of the endothelial lining of all blood vessels, and perhaps cells active in terms of pinecytosis of the extravascular compartment. This is a nonspecific stage, and all circulating proteins will be captured. FcRn has a very low affinity for albumin (and IgG) at serum pH, approximately pH 7.4.

2. Once subjected to pinocytosis, the vesicle formed is acidified to pH 5.0. Under acidic conditions, FcRn has a greater affinity for albumin, and binds to albumin, and also to IgG. Albumin and IgG thus bind to the FcRn receptor. FcRn binds to human serum albumin at a site in Domain III, through a site other than that by which it binds to IgG.

5 3. An ordering episode occurs, whereby most non-receptor-bound proteins are ordered in an endosome, in which most proteins will be directed for degradation. Albumin and IgG bound to the receptor are ordered in a vesicle directed to the cell surface, and thus prevent degradation.

4. The vesicle directed to the cell surface fuses with the cell surface, or briefly fuses with the cell surface.

cellular membrane. Under these conditions, the pH of the endosome increases to approach pH 7.4, the affinity of FcRn 10 for albumin is reduced, and albumin is released back into the circulation.

A clear set of desirable parameters can therefore be defined so that any SA binding protein has a maximum half-life. These parameters can be clearly illustrated using the FcRn serum albumin salvage receptor as a model, although it will also be applied to other receptors that mediate half-life prolongation.

15 • The affinity of the serum albumin binding will preferably be such that the SA binding protein does not dissociate from the albumin while experiencing glomerular filtration in the kidney, thus minimizing loss in urine, and / or

• The binding to SA will preferably not have a detrimental effect on the binding of serum albumin to any receptor responsible for maintaining serum albumin levels in the circulation, as it would inhibit recycling and,

20 with this, it would reduce the half-life of serum albumin and SA binder. Thus, SA binding dAb binds preferentially to an epitope other than that to which they bind by FcRn in domain III HSA, and the SA / dAb complex is preferably capable of coupling to FcRn, and / or

• SA binding will preferably be maintained under the conditions in which the receptor and the binder complex

United SA / SA are ordered or recycled. It has been shown that the endosomal pH approaches pH 5.0, and therefore a stable binding of dAb to serum albumin for pH 7.4 and pH 5.0 is desirable.

As illustrated in Example 15 below, most dAb bind to the 2nd HSA domain and therefore are not expected to compete with the binding of human serum albumin to FcRn. Two dAb (DOM7h-27 and DOM7h-30) join Domain III.

An anti-SA dAb that retains sufficient affinity for SA at a pH range of 7.4 to 5.0.

In addition to SA affinity, as well as in the absence of competition with the formation of SA: FcRn complexes, serum albumin specific dAb will preferably maintain SA affinity within a pH range of pH 7.4 in serum at pH 5 , 0 in the endosome to obtain full use of the rescue pathway mediated by FcRn.

In this pH range, only histidine residues and amino acid side chains with disturbed pKa

35 are likely to change protonation status. If the amino acid side chains make an important contribution to the binding energy of the complex, a shift in the pH from one end to the other in the range could be expected to produce a reduction in the binding affinity of the complex. Without wishing to be bound by any theory, this would in turn produce an increase in the probability that the specific SA dAb would enter the degradation pathway rather than be rescued by the FcRn-mediated salvage pathway.

Thus, for a binding to SA AlbudAb ™ (a dAb that specifically binds to serum albumin), it is desirable to select one in which the serum albumin binding characteristic does not change significantly with the pH (in the range of 5.0 to 7.4). A simple method of guaranteeing it would be to analyze the amino acid sequences of the dAb anti-SA regarding the absence of histidine residues in the CDRs. As shown below, several selection procedures for such property can be contemplated:

For example, a first round of selection is performed with the phage repertoire of ‘new’ phages using immobilized human serum albumin under conditions where the buffer pH is at pH 7.4 (eg PBS). The recovered and amplified phage population is then sent to a second selection run in which the incubation buffer is at pH 5.0. The alternation of buffers and pH is optionally repeated in additional batches in order to maintain the selection pressure for the binding of dAb to HSA for the two pHs.

50 In a second example, all selection batches are performed with the 'new' dAb phage repertoire using immobilized human serum albumin under conditions where the pH of the buffer is pH 7.4 (eg PBS) . However, just after washing the unbound phage with PBS (or PBS supplemented with Tween) and before elution of the bound phage, an additional wash / incubation step is added at pH 5.0 for a prolonged period of time ( eg up to 4 hours). During this period, phage presentation dAb that are unable to

55 binding to SA at pH 5.0 (but capable of binding to pH 7.4) is released from immobilized SA. After a second

series of washing steps at pH 5.0 with (without) Tween, the bound phage is recovered and analyzed.

In a third example, all selection batches are performed with the phage repertoire of dnew ’using human immobilized serum albumin under conditions where the pH of the buffer is pH 7.4 (eg PBS). Next, the best dAb candidates (ie, capable of binding at pH 7.4 and pH 5.0) are identified by screening. Normally, the genes encoding the dAb are recovered from the phage selected from the pool, subcloned into an expression vector that directs the soluble dAb to the culture supernatant of E. coli. Individual clones are collected, microtiter plate wells are cultured separately and induced for expression. The supernatants (or purified dAb) are then loaded directly onto a Biacore chip to identify dAb that exhibit affinity for immobilized serum albumin. Each supernatant is screened for

10 binding (mainly that of dissociation of the sensorgram) with HSA under conditions where the buffer ‘in progress’ is at pH 7.4 or at pH 5.0. It should be noted that screening for dAb binding in Biacore would also be used as a preferred method to identify the best values of the two previous examples.

A ligand comprising an individual variable domain as defined herein is described herein, in which the individual variable domain specifically binds to serum albumin at a natural serum pH and at an intracellular vesicle pH. The natural serum pH is about 7 (e.g., 7.4), and wherein said intracellular vesicle pH may be between about 4.8 and 5.2, or it may be at a pH of about 5. In one embodiment, the individual variable domain can specifically bind to serum albumin with a pH between about 7 and 5, or it can be at a pH of 7.4. Without wishing to be bound by any theory, an additional feature of this ligand is that its individual variable domain that specifically binds to serum albumin does not substantially dissociate from serum albumin while undergoing glomerular filtration in the kidney. Without wishing to be bound by any theory, an additional feature of this ligand is that its individual variable domain that specifically binds to serum albumin does not substantially interfere with the binding of FcRn to serum albumin. This individual variable domain can be an individual variable domain of antibodies; the individual variable domain of antibodies can be a VH3 domain and / or the individual variable domain of antibodies can be a V kappa domain. This individual variable domain may comprise a non-immunoglobulin framework, e.g. eg, frameworks of CTLA-4, lipocalin, SpA, Affibody ™, GroEL, Avimer ™ or fibronectin, and may contain one or more of CDR1, CDR2 and / or CDR3 of an individual variable domain of antibodies that preferably, although not necessarily , specifically binds to serum albumin. The individual variable domain (s) of this ligand can specifically bind human serum albumin, and / or include serum albumin of one or more species, e.g. eg, a human host, rat, monkey, pig, rabbit, hamster, mouse or goat. The intracellular compartment may be any intracellular compartment of any cell of any animal, which includes an endosomal compartment or intracellular vesicle or a budding vesicle. The endosomal compartment can have a pH of about 5, or 5.0. The ligands described herein may contain one or more individual variable domains that include immunoglobulin and / or immunoglobulin domains in which the binding of serum albumin to the individual variable domain does not substantially competitively inhibit the binding of FcRn to serum albumin. These one or more individual variable domains may preferably specifically bind to serum albumin with a lower equilibrium dissociation constant (Kd).

or equal to 300 nM.

A method of selecting a ligand comprising a variable domain is described herein.

40, which contains an individual variable domain (monomer), or more than one individual variable domain (eg, multimer, fusion protein, conjugate and double specific ligand as defined herein) that is specifically linked to serum albumin, in which the individual variable domain specifically binds to human serum albumin at a natural serum pH, and in which the individual variable domain does not competitively inhibit the binding of human serum albumin to FcRn, and in which the variable domain individual specifically joins

Human serum albumin at a pH of an intracellular compartment, comprising the steps of: (A) selection of ligands comprising an individual variable domain that does not bind to regions of human serum albumin that bind FcRn, (B) from the ligands selected in step (A), selection of ligands comprising an individual variable domain that binds to serum albumin at said natural serum pH. (C) selection of the ligands selected in step (B) for which they bind to serum albumin at the pH of said compartment

50 intracellular. Alternatively, steps (A) and (B) can be reversed as follows: (A) selection of ligands comprising an individual variable domain that binds human serum albumin at said natural serum pH, (B) from the selected ligands in (A), selection of ligands comprising an individual variable domain that binds to human serum albumin outside of the HSA regions that bind FcRn, and (C) of the ligands selected in step (B), selection of those that bind to serum albumin at said pH of said

55 intracellular compartment. A method for selecting a ligand comprising an individual variable domain is also described, in which the individual variable domain specifically binds to human serum albumin at a natural serum pH, wherein the individual variable domain does not competitively inhibit serum albumin binding. human to FcRn, and in which the individual variable domain specifically binds to human serum albumin at a pH of an intracellular compartment, comprising the steps of: (A) selection of ligands comprising a

60 individual variable domain that does not bind to human serum albumin regions that bind FcRn, (B) from step (A) selection of ligands comprising an individual variable domain that binds serum albumin at said pH natural serum, and (C) genetic modification of the individual variable domain of step (B) so that it binds to serum albumin at said pH of said intracellular compartment. Alternatively the stages

(A) and (B) can be reversed as follows: (A) selection of ligands comprising an individual variable domain that binds to serum albumin at said natural serum pH, (B) from the ligands selected in (A) , selection of ligands comprising an individual variable domain that does not bind to human serum albumin regions that bind FcRn, and (C) genetic modification of the individual variable domain of step (B) of

5 so that it binds to serum albumin at said pH of said intracellular compartment.

Assay to determine whether an individual variable domain does not competitively inhibit the binding of human serum albumin to FcRn: a Biacore competition experiment can be used to determine whether an individual variable domain competitively inhibits the binding of serum albumin to an FcRn. An experimental protocol for this example is as follows. After coating a CM5 sensor chip (Biacore AB) at 25 ° C with approximately 1,100 resonance units (RU) of a purified FcRn at pH 7.4, human serum albumin (HSA) is injected onto the surface of the antigen in a single concentration (e.g., 1 µm) alone, and in combination with a series of dilution mixtures, each HSA mixture having increasing amounts of the individual variable domain in question. The RUs of the resulting binding are determined for the HSA alone and each of the HSA / individual variable domain mixtures. By comparing the united RUs of HSA alone with the united RUs of HSA + individual variable domain 15, it can be seen if FcRn competes with the individual variable domain to join HSA. If it competes, then when the concentration of individual variable domain in solution increases, the RUs of HSA bound to FcRn will decrease. If there is no competition, adding the individual variable domain will have no influence on how the HSA binds to FcRn. This competition assay may optionally be repeated at pH 5.0 for an individual variable domain that binds to HSA at pH 5.0 in order to determine if the individual variable domain inhibits

20 competitively binding serum albumin to FcRn for pH 5.0.

These ligands that have an individual variable domain, which contain an individual variable domain (monomer) or more than one individual variable domain (e.g., multimer, fusion protein, conjugate and double specific ligand as defined herein) which is specifically bound to serum albumin, in which the individual variable domain specifically binds to serum albumin at a natural serum pH and a vesicle pH

Intracellularly, it may further comprise at least one additional individual variable domain, in which each additional individual variable domain specifically binds to a non-serum albumin antigen at a natural serum pH, but does not bind the antigen at an intracellular vesicle pH. . The natural serum pH is about 7.4, and the pH of said intracellular vesicle is between about 4.8 and 5.2, and in some embodiments, the pH of said intracellular vesicle is about 5.

A method based on the above ideas includes the use of a bispecific binder with affinity for a serum albumin to prolong the half-life and an affinity to a desired target antigen, as described above, to direct a bound antigen for degradation, or recycling. As described above, a binding to a serum albumin residue is selected, so that the binding is of high affinity at pH 5.0, so that the molecule would be classified for non-degradation in the endosome by an FcRn-mediated process. . Then you

35 selects a desired target antigen binding moiety using a technique similar to that described above, with the proviso that, instead of selecting high affinity binding at pH 7.4 and pH 5 as described above, it is performed selection for high affinity binding at pH 7.4, and low or zero affinity for the target antigen at pH 5. One way to achieve this is to select residues with histidines on the contact surface. A fusion protein between the 2 molecules is mainly made by conventional techniques of

40 molecular biology, either by derivation and chemical crosslinking, or by genetic fusion. The result is an increase in the potency of a given AlbudAb ™ (a dAb that specifically binds to serum albumin) in vivo, designing an SA binding dAb that binds to SA at pH 5, while having a dAb participant that binds to a ligand, which has a low or zero affinity at pH 5. Without wanting to be limited by any theory, after endosomal recycling, the target molecule will be released, and will go to a degradation endosome and will degrade, while the AlbudAb ™

45 (a dAb that specifically binds to serum albumin) is recycled to bind to a new ligand by FcRn-mediated recycling. This method offers a fundamental advantage over PEGylated molecules or other half-life extension technologies, in which this pathway is not available for regeneration. Presumably in these cases, the bound ligand simply sits on the PEGylated moiety and occupies it.

A method of directing a degradation antigen comprising the

Administration of a ligand that has at least one individual variable domain, in which the individual variable domain specifically binds to serum albumin at a natural serum pH and an intracellular vesicle pH, and which also has at least one additional individual variable domain , wherein the individual variable domain specifically binds to an antigen other than serum albumin at a natural serum pH, but does not bind said antigen at an intracellular vesicle pH, thereby directing the antigen other than serum albumin for degradation. In the

Also described herein is a ligand that further comprises at least one additional individual variable domain, wherein said individual variable domain specifically binds to an antigen other than serum albumin at a natural serum pH, but does not bind said antigen to a Intracellular vesicle pH.

In vitro dAb selection in the presence of metabolites

In the ligands described herein is a ligand comprising a variable domain

60 individual, which contains an individual variable domain (monomer) or more than one individual variable domain (e.g., multimer, fusion protein, conjugate and double specific ligand as defined herein) that

it is specifically bound to serum albumin, in which the individual variable domain specifically binds to human serum albumin, and in which the specific binding of serum albumin by the individual variable domain is not essentially blocked by binding of drugs and / or metabolites and / or small molecules to one or more sites in serum albumin. The one or more sites in human serum albumin include Sudlow 1 site and Sudlow 2 site. The one or

5 more sites may be located in any combination of one or more human serum albumin domains selected from the group consisting of domain I, domain II and domain III.

The ligands described herein include a ligand comprising an individual variable domain, which contains an individual variable domain (monomer) or more than one individual variable domain (e.g., multimer, fusion protein, conjugate and ligand specific double as defined herein) that 10 is specifically bound to serum albumin, in which the individual variable domain specifically binds to human serum albumin, and in which the specific binding of serum albumin by said individual variable domain does not alters the binding characteristics of serum albumin for drugs and / or metabolites and / or small molecules bound to SA. In one embodiment, the individual variable domain of the ligand binds to serum albumin in the presence and / or absence of a drug, metabolite or other small molecule. And in another embodiment, the specific binding of serum albumin by said individual variable domain does not substantially alter the serum albumin binding characteristics for drugs and / or metabolites and / or small molecules bound to SA naturally in vivo, which includes, but preferably it is not limited to the drugs and / or metabolites and / or small molecules described in Fasano et al. (2005) 57 (12): 787-96. The extraordinary ligand binding properties of human serum albuminum, and Bertucci, C. et al. (2002) 9 (15): 1463-81, Reversible and covalent binding of drugs to human serum albumin: methodological

20 approaches and physiological relevance.

Drugs and / or metabolites and / or small molecules bound to SA may or may not overlap with drugs and / or metabolites and / or small molecules that do not substantially inhibit or compete with serum albumin for binding to the individual variable domain. The drugs and / or metabolites include, but are preferably not limited to warfarin, ibuprofen, vitamin B6, bilirubin theta, hemin, thyroxine, fatty acids, acetaldehyde, fatty acid metabolites, acyl glucuronide, bilirubin metabolites, halothane, salicylate , benzodapenes and 1-O-gemfibrocil-B-Dglucuronide. This inhibition or competition with serum albumin for individual variable domain binding by small molecules can occur both by direct displacement and by allosteric effects as described for changes induced by small molecule binding at the junction of other small molecules, see Ascenzi et to the. (2006) Mini Rev. Med. Chem. 6 (4): 483-9. Allosteric modulation of drug binding to human serum albumin, and 30 Ghuman J. et al. (2005) J. Mol. Biol. 353 (1): 38-52 Structural basis of the drug-binding to human serum albumin. In one embodiment, the small molecule, either alone or in concert with one or more other small molecules, and / or metabolites, and / or proteins and / or drugs, binds to serum albumin. In another embodiment, the small molecule either alone or in concert with one or more other small molecules, and / or metabolites, and / or proteins and / or drugs, does not substantially inhibit or compete with serum albumin for domain binding. individual variable In

Another embodiment, the small molecule, either alone or in concert with one or more other small molecules, and / or metabolites, and / or proteins and / or drugs, substantially inhibits or competes with serum albumin for domain binding individual variable

The individual variable domain can be an individual variable domain of antibodies. The individual variable domain of antibodies can be a VH3 domain. The individual variable domain of antibodies can be a V domain

40 kappa The individual variable domain may comprise one or more non-immunoglobulin frameworks. The non-immunoglobulin framework may include one or more of, but preferably is not limited to, CTLA-4, lipocalin, SpA, GroEL and fibronectin, and includes an Affibody ™ and an Avimer ™.

A method of selecting a single variable domain that binds to serum albumin is described herein, comprising the selection of a first variable domain for its ability to bind to serum albumin in the presence of one or more metabolites and / or drugs, in which the selection is made in the presence of one or more metabolites and / or drugs. Also described herein is a method for producing a specific double ligand comprising a first individual variable domain of immunoglobulins having a first binding specificity for serum albumin in the presence of a metabolite and / or drug, and a second individual variable domain. of immunoglobulins having a second binding specificity, the method comprising the steps of: (a) selection of a first variable domain for its ability to bind to a first epitope in the presence of one or more metabolites and / or drugs, (b ) selection of a second variable domain for its ability to bind to a second epitope, (c) combination of the variable domains; and (d) ligand selection for its ability to bind to serum albumin in the presence of said one or more metabolites and / or ligands and said second epitopes. This method may also include a step in which the first variable domain is selected for binding to said first epitope in the absence of a complementary variable domain, and / or in which the first variable domain is selected for binding to said first epitope in the presence of a third complementary variable domain in which said third variable domain is different from said second variable domain. These selection steps can be performed in the presence of a mixture of metabolites and / or drugs and / or proteins and / or small molecules. The selection steps can also be carried out as follows: (a) selection of 60 individual variable domains that bind to serum albumin in the presence of a first metabolite and / or drug and / or small molecule; and (b) from the domains selected in step (a), a domain is selected in the presence of a second metabolite and / or drug and / or small molecule. A method for producing a specific double ligand having a first individual variable domain of immunoglobulins having

a first binding specificity for serum albumin in the presence of one of between metabolite and / or drug and / or small molecule, and a second individual variable domain of immunoglobulins having a second binding specificity, the method having the steps of: (a ) selection of first variable domains for their ability to bind to serum albumin in the presence of one or more metabolites and / or drugs and / or small molecules, (b) selection of second variable domains for their ability to bind to an epitope, ( c) combination of the variable domains to provide ligands comprising a first and a second variable domain; Y

(d) from the ligands provided by step (c), and selection of a ligand for its ability to bind to serum albumin in the presence of one or more metabolites and / or drugs and its ability to bind said epitopes, producing thus a specific double ligand. In one embodiment, the first variable domain is selected for serum albumin binding in the absence of a complementary variable domain. In another embodiment, the first variable domain is selected for binding to the first epitope in the presence of a complementary variable domain in which the complementary variable domain is different from the second variable domain.

A ligand comprising an individual variable domain is described herein, in which the individual variable domain specifically binds to serum albumin in vitro at pH 7 and at a pH of the intracellular compartment,

15 and in which the individual variable domain is an individual variable domain of unnatural occurrence. Also described herein is a ligand comprising an individual variable domain of antibodies, in which the individual variable domain of antibodies specifically binds to serum albumin in vitro at pH 7 and at a pH of the intracellular compartment. In one embodiment the pH of the intracellular compartment is between 4.8 and 5.2. In another embodiment, the binding of serum albumin to the individual variable domain of antibodies does not substantially inhibit the binding of FcRn to serum albumin, as determined by an in vitro surface plasmon resonance proficiency test.

In another embodiment, the individual variable domain of antibodies comprises one or more CDR regions of antibodies selected from the group consisting of: CDR1, CDR2 and CDR3. In another embodiment, the individual variable domain of antibodies comprises a framework selected from the group consisting of: CTLA-4, lipocalin,

Staphylococcal protein A (SpA), GroEL, GroES, transferrin and fibronectin. The binding of serum albumin to the individual variable domain does not substantially compete with the binding of FcRn to serum albumin in one embodiment, and in another embodiment the individual variable domain of antibodies specifically binds to serum albumin with an equilibrium dissociation constant (Kd) less than or equal to 300 nM.

In another embodiment, the individual variable domain of antibodies further comprises at least one additional individual variable domain of antibodies, in which the additional individual variable domain of antibodies specifically binds to an antigen other than serum albumin at pH 7, but does not bind to the antigen at the pH of the intracellular compartment. Also described herein is a method of addressing an antigen for degradation in a person comprising the administration of a ligand comprising an individual variable domain, such as an individual variable antibody domain, which specifically binds to

In vitro serum albumin at pH 7 and at a pH of the intracellular compartment, for the individual, the ligand further comprising at least one additional individual variable domain of antibodies comprising an individual variable domain, e.g. eg, an individual variable domain of antibodies, in which the antigen other than serum albumin is the antigen targeted for degradation.

In one embodiment of the ligands of the invention, the specific binding of human serum albumin by the individual variable domain of antibodies is not blocked by the binding of a predetermined drug and / or metabolite and / or a small molecule to one or more sites. in human serum albumin. In these embodiments, the additional individual variable domain of antibodies may be an individual variable heavy chain antibody domain or an individual variable light chain domain of antibodies comprising one or more CDRs of antibodies selected from the group consisting of: CDR1, CDR2 and / or CDR3. The individual variable domains

45 may comprise a framework selected from the group consisting of: CTLA-4, lipocalin, staphylococcal protein A (SpA), GroEL, GroES, transferrin and fibronectin.

Another embodiment of a ligand described herein is a ligand comprising an individual variable domain, in which the individual variable domain is an individual variable domain of unnatural occurrence, in which the individual variable domain specifically binds to serum albumin. human in vitro at pH 7 and at a pH of the intracellular compartment, in which the specific binding of human serum albumin by the individual variable domain is not blocked by the binding of a drug and / or a metabolite and / or a predetermined small molecule to one or more sites in human serum albumin, wherein the one or more sites in human serum albumin include Sudlow 1 site and Sudlow 2 site or the one or more sites are located in one or more domains of human serum albumin selected from the group consisting of: domain I, domain II and domain III.

55 Linker Groups

The connection of an AlbudAb ™ (a dAb that specifically binds to serum albumin) (serum anti-albumin domain antibody or individual variable domain) with another biologically active moiety can be obtained by means of recombinant engineering techniques. Basically, the genes that encode the two proteins of interest are fused in the frame. Several formats can be considered in which the serum albumin domain antibody is at the N-terminus of the fusion (ie AlbudAb ™ -Y in which Y is a biologically active polypeptide), in the

C-terminus of the fusion (ie Z-AlbudAb ™ in which Z is a biologically active peptide). In some cases, the fusion of more than one biologically active polypeptide into an AlbudAb ™ (a dAb that specifically binds to serum albumin) can be considered, resulting in a series of possibilities related to the design of the fusion. For example, the fusion could be as follows: Z-Y-AlbudAb ™, Z-AlbudAb ™ -Y or AlbudAb ™ -Z-Y.

In all these fusion molecules, two polypeptides are covalently linked through at least one peptide bond. In its simplest approach, AlbudAb ™ (a dAb that specifically binds to serum albumin) and the biologically polypeptide (s) are directly linked. Thus, the binding between AlbudAb ™ (a dAb that specifically binds to serum albumin) and the polypeptide would be as follows:

a) For an AlbudAb ™ (a dAb that specifically binds to serum albumin) at the C-terminus,

In which AlbudAb ™ is a VK:

xxxDIQ

xxxNIQ

xxxAIQ

xxxAIR

xxxVIW

xxxDIV

xxxDVV

xxxEIV

xxxETT

In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a Vλ:

xxxQSV

xxxQSA

xxxSYE

xxxSSE

xxxSYV

xxxLPV

xxxQPV

xxxQLV

xxxQAV

xxxNFM

xxxQTV

xxxQAG

In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a VH (e.g., human VH):

xxxQVQ

xxxQMQ

xxxEVQ

xxxQIT

xxxQVT

xxxQLQ

In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a VHH (e.g., camelid heavy chain variable domain): xxxEVQ xxxQVQ 5 xxxDVQ xxxQVK xxxAVQ b) For an AlbudAb ™ (a dAb which specifically binds to serum albumin) at the N-terminus, in which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a VK: 10 KVEIKxxx KLEIKxxx KVDIKxxx RLEIKxxx EIKRxxx 15 In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a Vλ: -KVDVLxxx KLDVLxxx QLDVLxxx In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a VH (e.g., human VH):

20 VTVSSxxx In which AlbudAb ™ (a dAb that specifically binds to serum albumin) is a VHH (e.g., camelid heavy chain variable domain): -

VTVSSxxx ’xxx’ represents the first three or last three amino acids of the biologically active (first) polypeptide fused with AlbudAb ™ (a dAb that specifically binds to serum albumin). However, there may be cases in which the production of a recombinant fusion protein that recovers the functional activity of the two polypeptides can be facilitated by the connection to the coding genes with a bridge DNA segment encoding a peptide linker group that is divide among tandem connected polypeptides. The optimal length of the peptide linker group is usually empirically conceived: it can be very short as an amino acid or extend up to 50 amino acids. Linker groups of different designs have been proposed and are well known in the art. It is understood that the following examples provide an extensive, but not exhaustive, list of possible linker group approaches:

1. Flexible linking groups:

Flexible linker groups are designed to adopt a non-stable secondary structure when

35 connect two polypeptide moieties, thus allowing a range of conformations in the fusion protein. These linker groups are preferably hydrophilic in nature to prevent them from interacting with one or both fused polypeptides. Normally in these linker groups, small polar residues such as glycine and serine dominate in the linker groups in order to increase the flexible and hydrophilic characteristics of the peptide's main structure, respectively. The length of these linker groups is variable and is

40 determines better empirically or with the help of 3D calculation strategies. In general, a preferred linker group length will be the minimum compatible with good expression, good solubility and full recovery of the native functions and structures of interest. Due to their flexible characteristics, flexible linker groups can be good substrates for endogenous proteases. In general, unless a desirable feature is flexible linker groups are devoid of amino acids such as charged amino acids or

45 large hydrophobic / aromatic ones that are efficiently recognized by endogenous proteases with a broad substrate specificity. In addition cysteine residues are preferably avoided since free cysteines

they can react with each other to form cysteines, from which (i) the formation of bridges between two fusion proteins occurs through the linker groups, and / or (ii) the compromise of the expression / folding of the fusion protein if one or more of the bioactive polypeptides comprise one or more cysteine residues ('cysteine randomization').

5 Examples of flexible linker groups are: (i) glycine-rich linker groups based on the repetition of a motif (GGGGS) and in which y is at least 1, although and may be 2, 3, 4, 5, 6 , 7, 8 and 9, or more (see PCT International Publications No. EP 0 753 551, US 5 258 498, EP 0 623 679), (ii) serine-rich linker groups based on the repetition of a motif (SSSSG) and in which y is at least 1, although and may be 2, 3, 4, 5, 6, 7, 8 and 9, or more (see PCT international publications no .: EP 0 573 551, US 5 525 491).

10 2. Restricted linker groups:

Restricted linker groups are designed to adopt a stable secondary structure when two polypeptide moieties are connected, thereby restricting the range of conformations in the fusion protein. Such linker groups normally adopt a helical structure that extends several turns. Again, the length of these linker groups is variable and is best determined empirically or with the help of calculation strategies.

15 The main reason for choosing restricted linker groups is to maintain maximum distance between each fusion polypeptide. This fact is especially relevant when the two polypeptides have a tendency to form heteroaggregates. Due to their structure, restricted linker groups may also be more resistant to proteolytic degradation, thereby offering an advantage when injected in vivo.

Examples of restricted linker groups are cited in PCT International Publications No. WO-00/24884

20 (eg SSSASASSA, GSPGSPG, or ATTTGSSPGPT), US 6,132,992 (eg helical peptide linker groups).

3. ’Natural’ linking groups:

Natural linker groups are polypeptide sequences (of varying lengths) that, by contrast, are not synthetic, that is, the polypeptide sequences that make up the linker groups are present in nature. Natural linker groups can be flexible or restricted and can be very diverse in amino acid sequence and composition. Their degree of resistance to proteolysis depends on the proteins of origin and the biological environment to which these proteins are exposed in nature (extracellular, intracellular, prokaryotic, eukaryotic, etc.). A class of linker groups is especially relevant for the development of biological therapeutic products in humans: linker groups based on peptides present in

30 human proteins In fact, these linker groups are by nature non-immunogenic, or are very weak, and therefore potentially safer for human therapy.

Examples of natural linker groups are: (i) KESGSVSSEQLAQFRSLD (see Bird et al. (1988) Science, 242, 423-426), (ii) sequences corresponding to the hinge domain of immunoglobulins devoid of light chains (see Hamers-Casterman et al. (1993) Nature, 363, 446-448 and PCT International Publication No.: WO35 096/34103). Examples of linker groups for use with anti-albumin domain antibodies (e.g., human, humanized, human with camelid or VHH domain antibodies of camelids) are EPKIPQPQPKPQPQPQPQPKPQPKPEPECTCPKCP and GTNEVCKCPKCP. Other linker groups obtained from human hinges and camelids are described in EP0656946, incorporated herein by reference. Linker groups derived from hinges can have variable lengths, for example 0

40 to about 50 amino acids, which includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 11, 12,13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48 or 49 amino acids.

As used herein, "drug" refers to any compound (eg, small organic molecule, nucleic acid, polypeptide) that can be administered to an individual to produce a beneficial effect,

Therapeutic or diagnostic, through binding to and / or altering the function of a biological target molecule in the individual. The target molecule may be an endogenous target molecule encoded by the genome of the individual (eg, an enzyme, receptor, growth factor, cytokine encoded by the genome of the individual) or an exogenous target molecule encoded by the genome of a pathogen ( eg an enzyme encoded by the genome of a virus, bacteria, fungus, nematode or other pathogen).

The composition of the drug can be a conjugate wherein the drug is covalently or non-covalently bound to the rest of the polypeptide binding. The drug may be covalently or non-covalently bound to the polypeptide binding moiety directly or indirectly (e.g., through a suitable linker group and / or non-covalent binding of complementary binding partners (e.g., biotin and avidin)). When complementary binding participants are used, one of the binding partners can be covalently linked to the drug directly or to

Through a residue of the suitable linker group, and the complementary binding partner may be covalently linked to the polypeptide binding residue directly or through a suitable linker group residue. When the drug is a polypeptide or peptide, the composition of the drug may be a fusion protein, wherein the polypeptide or peptide, the drug and the polypeptide binding moiety are discrete parts (residues) of a

continuous polypeptide chain. As described herein, polypeptide binding moieties and polypeptide drug moieties may be linked directly to each other through a peptide bond, or linked through an appropriate amino acid, or peptide or polypeptide linker group.

Decrease of immunogenicity

5 A method of reducing the immunogenicity of a pharmaceutical agent is described herein, which comprises modifying said agent so that the agent further contains a region of an individual variable domain, in which the individual variable domain specifically binds to serum albumin in vivo and / or ex vivo, and in which the agent may include a drug, a metabolite, a ligand, an antigen and a protein. The serum albumin can be human serum albumin. The individual variable domain can be a variable domain

10 individual immunoglobulins. The individual variable domain of immunoglobulins can be an individual variable domain VH of antibodies. The individual variable domain VH can be an individual variable domain VH3. The individual variable domain VH3 can be an individual variable domain VH3 human. The individual variable domain can be an individual Vkappa or Vlambda antibody variable domain. The individual variable domain of antibodies may comprise a set of four Kabat framework regions (FRs that are encoded by

15 gene segments of germline antibodies of frame VH3). The VH3 frame is selected from the group consisting of DP47, DP38 and DP45. The individual variable domain of antibodies may contain a set of four Kabat framework regions (FR), which are encoded by gene segments of VKappa framework germline antibodies. A non-limiting example of a Kappa framework is DPK9. The individual variable domain may contain an immunoglobulin or non-immunoglobulin framework containing CDR1, CDR2 and / or CDR3 regions, in

20 where at least one of the CDR1, CDR2 and CDR3 regions is of a variable antibody domain that specifically binds to serum albumin. The non-immunoglobulin framework may include, but is preferably not limited to, CTLA-4, lipocalin, SpA, Affibody ™, GroEL, Avimers ™ and fibronectin. The serum albumin can be human serum albumin. The individual variable domain of immunoglobulins and / or the individual variable domain of non-immunoglobulins can specifically bind human serum albumin with a Kd of less than 300 nM. He

The individual variable domain of immunoglobulins and / or the individual variable domain of non-immunoglobulins can specifically bind human serum albumin and one or more non-human serum albumin, with Kd values within the limits of a factor of 10 to each other. The individual variable domain of immunoglobulins and / or the individual variable domain of non-immunoglobulins can specifically bind human serum albumin and one or more non-human serum albumin, and wherein the beta T half-life of the ligand is substantially the same as the

30 beta half-life of human serum albumin in a human host. In addition, the individual variable domain of immunoglobulins and / or the individual variable domain of non-immunoglobulins can specifically bind Domain II of human serum albumin. The individual variable domain of immunoglobulins and / or the individual variable domain of non-immunoglobulins can specifically bind to serum albumin at a natural serum pH and at an intracellular vesicle pH. The specific binding of serum albumin by said individual variable domain of

Immunoglobulins and / or individual variable domain of non-immunoglobulins are preferably not substantially blocked by drug and / or metabolite binding to one or more sites in serum albumin. In one embodiment, the specific binding of serum albumin by the individual variable domain does not alter the binding characteristics of serum albumin for drugs and / or metabolites and / or small molecules bound to SA. In one embodiment the method for modifying the agent results in the formation of a modified agent that has a formula that

40 comprises: a- (X) n1-b- (Y) n2-c- (Z) n3-d or a- (Z) n3-b- (Y) n2-c- (X) nd, where X it is a polypeptide drug that has binding specificity for a first target; And it is an individual variable domain, e.g. ex. an individual variable domain of antibodies that specifically binds to serum albumin in vivo and / or ex vivo; Z is a polypeptide drug that has binding specificity for a second target; a, b, c and d are independently a polypeptide comprising from one to about 10 amino acid residues or is absent; n1 is from one to

45 about 10; n2 is from one to about 10; and n3 is zero to about 10. In a further embodiment, when n1 and n2 are one and n3 is zero, X does not comprise any antibody chain or any antibody chain fragment.

A method of reducing the immunogenicity of a pharmaceutical agent is described herein, which comprises modifying the agent so that the agent further comprises an individual variable domain, in which the individual variable domain specifically binds to serum albumin, in which the individual variable domain is an individual variable domain of unnatural occurrence, and in which the agent is selected from the group consisting of: a drug, a metabolite, a ligand, an antigen and a protein. Also described herein is a method for reducing the immunogenicity of a pharmaceutical agent, which comprises modifying the agent so that the agent further comprises an individual variable domain of antibodies, in which the individual variable domain of antibodies binds. specifically to serum albumin, and in which the agent is selected from the group consisting of: a drug, a metabolite, a ligand, an antigen and a protein. In one embodiment, the individual variable domain of antibodies is an individual heavy chain variable domain of antibodies, e.g. eg, individual variable domain VH3 of antibodies, or an individual variable domain VH3 of human antibodies. In another embodiment, the individual variable domain of antibodies is a single variable domain of the light chain of antibodies, e.g. eg, an individual variable domain Vkappa of antibodies or Vlambda of antibodies. In one embodiment, the individual variable antibody domain comprises CDR1, CDR2 and CDR3 regions, in which at least one of the CDR1, CDR2 and CDR3 regions is of a variable antibody domain that specifically binds to serum albumin, and optionally comprises also a frame selected from

group consisting of: CTLA-4, lipocalin, staphylococcal protein A (SpA), GroEL, GroES, transferrin and fibronectin. In another embodiment of these methods, the individual variable domain, e.g. For example, the individual variable domain of antibodies is specifically bound to human serum albumin with a Kd of less than 300 nM, and in another embodiment of these methods, the individual variable domain, e.g. eg, the individual variable domain of antibodies, is

5 specifically bound to human serum albumin and one or more non-human serum albumin, with Kd values in the limits of a factor of 10 to each other. In another embodiment of these methods, the individual variable domain, e.g. For example, the individual variable domain of antibodies is specifically bound to human serum albumin and non-human serum albumin, and the beta T half-life of the ligand is substantially the same as the human serum albumin beta T half-life in a human host. In another embodiment of these methods, the individual variable domain, e.g. eg, the individual variable domain of antibodies, is specifically bound to a domain II of human serum albumin. In another embodiment of these methods, the individual variable domain, e.g. eg, the individual variable domain of antibodies, specifically binds to serum albumin at a pH 7 and at a pH of the intracellular compartment.

The invention is further described, for illustration purposes only, in the following examples. As used herein, for the purposes of naming dAb, the human TNFα receptor is referred to as TAR1 and

The human TNFα receptor 1 (p55 receptor) is referred to as TAR2.

Example 9

Selection of a collection of individual antibody domains (dAb) directed against human serum albumin (HSA) and mouse serum albumin (MSA).

This example explains a method for preparing a single domain antibody (dAb) directed against serum albumin. The selection of dAb against mouse serum albumin (MSA) and human serum albumin (HSA) is described. In this experiment, three human antibody phage display libraries were used, each based on an individual human framework for VH (see Figure 13: VH training sequence based on V3-23 / DP47 and JH4b) or Vκ (see Figure 15 : training Vκ sequence based on o12 / o2 / DPK9 and Jk1) with diversity of side chains encoded by NNK codons incorporated in complementarity determination regions

25 (CDR1, CDR2 and CDR3).

Library 1 (VH):

Diversity in positions: H30, H31, H33, H35, H50, H52, H52a, H53, H55, H56, H58, H95, H97, H98.

Library size: 6.2 x 109

Library 2 (VH):

Diversity in positions: H30, H31, H33, H35, H50, H52, H52a, H53, H55, H56, H58, H95, H97, H98, H99, H100, H100a, H100b.

Library size: 4.3 x 109

Library 3 (Vκ):

Diversity in positions: L30, L31, L32, L34, L50, L53, L91, L92, L93, L94, L96

35 Library size: 2 x 109

The VH and Vκ libraries have been preselected for binding to protein A and protein L of generic ligands respectively so that most of the clones in the unselected libraries are functional. The sizes of the libraries shown above correspond to the sizes after the preselection.

Two batches of serum albumin were made using each of the libraries separately. For each selection, the antigen was coated on an immunotube (nunc) in 4 ml of PBS at a concentration of 100 µg / ml. In the first round of selection, each of the three libraries was analyzed separately in relation to HSA (Sigma) and MSA (Sigma). In the second round of selection, a phage of each of the six selections of the first round was analyzed in relation to (i) again the same antigen (eg MSA of 1st round, MSA of 2nd round) and (ii) in relation to the reciprocal antigen (eg MSA of 1st round, HSA of 2nd round) to produce a total of

45 twelve selections from the 2nd round. In each case, after the second round of selection 48 clones were tested for binding to HSA and MSA. Soluble dAb fragments were produced as described for the scFv fragment in Harrison et al., Methods Enzymol. nineteen ninety six; 267: 83-109 and the standard ELISA protocol (Hoogenboom et al. (1991) Nucleic Acid Res., 19: 4133) was followed with the proviso that 2% Tween PBS was used as blocking buffer and the bound dAb were detected with L-HRP protein (Sigma) (for Vκ) and A-HRP protein (Amersham Pharmacia Biotech) (for VH).

The dAb that produced a signal above the background indicating binding to MSA, HSA or both were tested in an insoluble form of ELISA for plastic binding alone but all were specific for serum albumin. The clones were then sequenced (see Table 5) which highlighted that they had been identified

21 unique dAb sequences. The minimum similarity (at the amino acid level) among the selected Vκ dAb clones was 86.25% ((69/80) x 100; the result when all diversified residues are different, eg clones 24 and 34). The minimum similarity among the selected VH dAb clones was 94% ((127/136) x 100).

5 Next, the serum albumin binding dAb were tested for their ability to capture biotinylated antigen from the solution. The ELISA protocol (as before) was followed with the proviso that ELISA plates were coated with 1 µg / ml of L proteins (for Vκ clones) and 1 µg / ml of A proteins (for VH clones) . The soluble dAb was captured from the solution as in the protocol and detection was performed with biotinylated MSA or HSA and streptavidin HRP. Biotinylated MSA and HSA have been prepared following the instructions

10 from the manufacturer, in order to get an average of 2 biotins per serum albumin molecule. Twenty-four clones were identified that captured biotinylated MSA of the solution in ELISA, Table 5. Two of them (clones 2 and 38 shown below) also captured biotinylated HSA. Next, the dAb were tested for their ability to bind MSA coated on a Biacore CM5 chip. Eight clones were found that joined the MSA in Biacore.

15 Table 5

dAb (all capture biotinylated MSA)

H or κCDR1 CDR2 CDR3 Joins MSA in Biacore BSA capture biotinylated

Vκ 3 library mold (training)

κ XXXLX XASXLQS QQXXXXPXT

2, 4, 7, 41,

κ SSYLN RASPLQS QQTYSVPPT ✓ the 4 join

38, 54

κ SSYLN RASPLQS QQTYRIPPT ✓ the two come together

46, 47, 52, 56

κ FKSLK NASYLQS QQVVYWPVT

13, 15

κ YYHLK KASTLQS QQVRKVPRT

30, 35

κ RRYLK QASVLQS QQGLYPPIT

19,

κ YNWLK RASSLQS QQNVVIPRT

22,

κ LWHLR HASLLQS QQSAVYPKT

2. 3,

κ FRYLA HASHLQS QQRLLYPKT

24,

κ FYHLA PASKLQS QQRARWPRT

31,

κ IWHLN RASRLQS QQVARVPRT

33,

κ YRYLR KASSLQS QQYVGYPRT

3. 4,

κ LKYLK NASHLQS QQTTYYPIT

53,

κ LRYLR KASWLQS QQVLYYPQT

eleven,

κ LRSLK AASRLQS QQVVYWPAT ✓

12,

κ FRHLK AASRLQS QQVALYPKT ✓

17,

κ RKYLR TASSLQS QQNLFWPRT ✓

18,

κ RRYLN AASSLQS QQMLFYPKT ✓

16, 21

κ IKHLK GASRLQS QQGARWPQT ✓

25.26

κ YYHLK KASTLQS QQVRKVPRT ✓

27,

κ YKHLK NASHLQS QQVGRYPKT ✓

55,

κ FKSLK NASYLQS QQVVYWPVT ✓

VH 1 (and 2) library mold (training)

H XXYXXX XIXXXGXXTXYADSVKG XXXX (XXXX) FDY

8, 10

H WVYQMD SISAFGAKTLYADSVKG LSGKFDY

36,

H WSYQMT SISSFGSSTLYADSVKG GRDHNYSLFDY

In all cases the frames were identical to the frames in the corresponding training sequence, with diversity in the CDRs as indicated in Table 5.

Of the eight clones that joined MSA in Biacore, two clones that are highly expressed in E. 5 coli (MSA16 and MSA26 clones) were chosen for further study (see Example 10). The complete nucleotide and amino acid sequences for MSA16 and 26 are shown in Figure 16.

Example 10

Determination of affinity and serum half-life in mice of MSA16 and MSA26 binding dAb.

As described in US20060251644, a common method for determining binding affinity

10 consists in assessing the association and dissociation rate constants using a Biacore ™ surface plasmon resonance system (Biacore, Inc.). A biosensors chip is activated for covalent coupling of the target following the manufacturer's instructions (Biacore). The target is then diluted and injected into the chip to obtain a signal in response units (RU) of immobilized material. Since the signal in RU is proportional to the mass of immobilized material, this represents a range of immobilized target densities in

15 the matrix. Dissociation data fit a model of a site to obtain koff +/- d.t. (standard deviation of measures). Velocity constants (Kd) of pseudo-first order are calculated for each association curve, and plotted as a function of protein concentration to obtain kon +/- e.s. (standard error of adjustment). The equilibrium dissociation constants for the junction, Kd, are calculated from SPR measurements such as koff / kon.

The dAb MSA16 and MSA26 were expressed in the periplasm of E. coli and purified using batch absorption for the affinity resin of the L-agarose protein (Affitech, Norway) followed by elution with glycine at pH 2.2. The purified dAb were then analyzed by Biacore inhibition to determine Kd. Briefly, the purified MSA16 and MSA26 were tested to determine the concentration of dAb needed to achieve 200 RU of response on a Biacore CM5 chip coated with a high density of MSA. Once determined the

At the required concentrations of dAb, the MSA antigen was premixed at a range of concentrations around the expected Kd with the dAb and incubated overnight. The binding to the BAcore chip coated with dAb MSA in each of the premixes at a high flow rate of 30 µl / minute was then measured. Affinities are determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991). The Biacore system (Uppsala, Sweden) is a preferred method of determining binding affinity. The Biacore system

30 uses surface plasmon resonance (SPR, Welford K. 1991, Opt. Quant. Elect. 23: 1; Morton and Myszka, 1998, Methods in Enzymology 295: 268) to control biomolecular interactions in real time. Biacore analysis conveniently generates association speed constants, dissociation rate constants, equilibrium dissociation constants and affinity constants. The resulting curves were used to create graphical representations of Klotz (Klotz, I.M. (1982) Science 217: 1247-1249 and Klotz, I.M. (1983) J. Trends in

35 Pharmacol. Sci. 4: 253-255) which provided an estimated Kd of 200 nM for MSA16 and 70 nM for MSA 26 (Figure 17A and B).

Next, MSA16 and MSA26 clones were cloned into an expression vector with the brand HA (nucleic acid sequence: TATCCTTATGATGTTCCTGATTATGCA and amino acid sequence: YPYDVPDYA) and amounts of 2-10 mg were expressed in E. coli and purified to from the affinity resin supernatant of the L-agarose protein (Affitech, Norway) and eluted with glycine at pH 2.2. The serum half-life of the dAb in mice was determined. MSA26 and MSA16 were dosed as i.v. injections. individual at approximately 1.5 mg / kg in CD1 mice. The analysis of serum levels was performed by capturing goat anti-HA (Abcam, RU) and detecting L-HRP (Invitrogen) proteins. The ELISA was blocked with 4% Marvel. Washing was performed with 0.05% PBS Tween. Standard curves of known concentrations of dAb in the presence of serum were determined

45 1x mouse to ensure comparability with test samples. Modeling with a 2-compartment model showed that MSA-26 had a t1 / 2α of 0.16 h, a t1 / 2β of 14.5 h and an area under the curve (ABC) of 465 h.mg/ml (data not shown) and MSA-16 had a t1 / 2α of 0.98 h, a t1 / 2β of 36.5 h and an ABC of 913 h.mg/ml (Figure 18). The two anti-MSA clones had a considerably longer half-life than with HEL4 (a chicken egg white antilisozyme dAb) that had a t1 / 2α of 0.06 h and a t1 / 2β of 0.34 h.

50 Example 11. Creation of double specific Fab fragments of VH-VH and Vκ-Vκ

This example describes a method for preparing specific double VH-VH and Vκ-Vκ as Fab fragments.

Before constructing each of the described Fab-type fragments, the dAb that bind to the targets of choice between dAb libraries similar to those described in example 9 were selected first. A VH dAb, HEL4, was isolated which is bound to chicken egg lysozyme (Sigma) and a second VH dAb (TAR2h-5) that is bound to a TNFα receptor (R and D systems) was also isolated. The sequences thereof are provided in the list.

5 sequences. A Vκ dAb that binds to TNFα (TAR1-5-19) was isolated by affinity selection and maturation and the sequence is also indicated in the sequence listing. In these experiments a second Vκ dAb (MSA 26) described in example 9 was also used, the sequence of which is in Figure 17B.

The DNA of the expression vectors containing the four dAb described above was digested with SalI and NotI enzymes to cleave the DNA encoding the dAb. A band of the expected size was purified (300-400 bp)

10 by passing the digested material on an agarose gel and cleaving the band, followed by gel-type purification using the Qiagen gel purification kit (Qiagen, RU). The DNA encoding the dAb was then inserted into the CH or Cκ vectors (Figures 8 and 9) as indicated in Table 6.

Table 6

dAb

Target antigen VH of dAb or Vκ of dAb Introduced in vector Mark (end C) Antibiotic resistance

HEL4

Chicken Egg Lysozyme Vh CH myc Chloramphenicol

TAR2-5

TNF receiver Vh Cκ flag Ampicillin

TAR1-5-19

TNF α Vκ CH myc Chloramphenicol

MSA 26

Mouse serum albumin Vκ Cκ flag Ampicillin

15 The CH constructs of VH and Cκ of VH were co-transformed into HB2151 cells. Separately, the Vκ CH and Vκ Cκ constructs were cotransformed into HB2151 cells. Cultures of each of the cotransformed cell lines were grown overnight (in 2xTy containing 5% glucose, 10 µg / ml chloramphenicol and 100 µg / ml ampicillin to maintain antibiotic selection for plasmids of CH and Cκ). The overnight cultures were used to inoculate new medium (2xTy, 10 g / ml of chloramphenicol and 100 g / ml

20 ampicillin) and were grown at OD 0.7-0.9 before induction by adding IPTG to express their CH and Cκ constructs. The Fab type fragment was then purified from the periplasm by purification of A proteins (for CH of VH and Cκ of VH cotransformed) and purification with affinity resin of MSA (for CH of Vκ and Cκ of Vκ cotransformed).

VH-VH specific double

The expression of CH of VH and Cκ of double specific VH was tested by applying the protein on a gel. The gel was transferred and a band of the expected size for the Fab fragment could be detected in the Western blot by means of the myc mark and the flag mark, indicating that the CH parts of VH and Cκ of VH of the Fab type fragment They were present. Next, in order to determine whether the two halves of the double specific were present in the same Fab type fragment, an ELISA plate was coated overnight at 4

30 ° C with 100 µl per well of chicken egg lysozyme (HEL) at 3 mg / ml in sodium bicarbonate buffer. The plate was then blocked (as described in Example 1) with 2% Tween PBS followed by incubation with the VH / VKC Fab CH fragment. The detection of HEL double specific binding was performed by means of the non-similar chain using 9e10 (a monoclonal antibody that binds to the myc brand, Roche) and anti-mouse IgG-HRP (Amersham Pharmacia Biotech). The signal for the VH / Cκ Fab CH type fragment

The specific double was 0.154 compared to a 0.069 background signal for the VH Cκ chain expressed alone. This demonstrates that the Fab type fragment has binding specificity for the target antigen.

Double specific Vκ-Vκ

After purifying the Fab type fragment of CH of Vκ and Cκ of specific Vκ double cotransformed into a

With the affinity resin of MSA, the resulting protein was used to probe an ELISA plate coated with 1 µg / ml TNFα and an ELISA plate coated with 10 µg / ml MSA. As predicted, there was a signal above the background when detected with L-HRP protein in the two ELISA plates (data not shown). This indicated that the protein fraction capable of binding to MSA (and therefore purified in the affinity column of MSA) was also capable of binding to TNFα in a subsequent ELISA, confirming the double specificity of the fragment of

45 antibody. This protein fraction was then used for two subsequent experiments. First, an ELISA plate coated with 1 µg / ml of TNFα was probed with Vκ and Cκ double specific Fab type fragment

of Vκ and also with a control TNFα binding dAb at a concentration calculated to provide a similar signal in ELISA. The double specific and control dAb were used to probe the ELISA plate in the presence and in the absence of 2 mg / ml of MSA. The signal in the double specific well was reduced by more than 50% but the signal in the dAb well was not reduced at all (see Figure 19a). The same protein was also placed in the

5 receptor with and without MSA and competition for MSA was shown (see Figure 19c). Thus it is demonstrated that the binding of MSA to the specific double is competitive with the binding to TNFα.

Example 12. Creation of a specific double Vκ-Vκ cis bound with specificity for mouse serum albumin and TNFα

This example describes a method for preparing a specific double antibody fragment specific for

10 mouse serum albumin and TNFα by chemical coupling via a disulfide bond. MSA16 dAb (from example 1) and TAR1-5-19 were cloned back into a pET-based vector with a cysteine at the C-terminus and no labels. The two dAb were expressed in levels of 4-10 mg and purified from the supernatant using the affinity resin of the L-agarose protein (Afitiech, Norway). The cysteine-labeled dAb were then reduced with dithiothreitol. The dAb of TAR1-5-19 was then coupled with dithiodipyridine to block the

Reform of disulfide bonds that produce the formation of PEP homodimers 1-5-19. The two different dAb were then mixed at pH 6.5 to promote the formation of disulfide bonds and the generation of cis-linked heterodimers TAR1-5-19, MSA16. This method of producing conjugates of two different proteins was originally described by King et al. (King TP, Li and Kochoumian L Biochemistry. 1978 vol 17: 1499-506 Preparation of protein conjugates via intermolecular disulfide bond formation). The heterodimers were separated from

20 monomeric species by cation exchange. The separation was confirmed by the presence of a band of the expected size in an SDS gel. The resulting heterodimeric species was tested in the TNF receptor assay and was found to have an IC50 to neutralize TNF of approximately 18 nM. Next, the receptor assay was repeated with a constant concentration of heterodimer (18 nM) and a dilution series of MSA and HSA. The presence of HSA in a concentration range (up to 2 mg / ml) did not produce a

25 reduction in the ability of the dimer to inhibit TNFα. However, the addition of MSA resulted in a dose-dependent reduction in the ability of the dimer to inhibit TNFα (Figure 20). This demonstrates that MSA and TNFα compete for binding in the TAR1-5-19 dimer, cis-bound MSA16 .

Data summary

Annex 4 shows a summary of data obtained in the experiments included in the preceding examples.

30 Example 13.

Selection and characterization of dAb for serum albumin binding from a set of species.

DAb versus human serum albumin, mouse serum albumin and swine serum albumin were selected as described above for the anti-MSA dAb except for the following modifications in the protocol: the VH synthetic domain phage libraries were the 4G and 6G libraries , which are based on a human VH3 comprising the DP47 germline gene and the JH4 segment for VH and a human Vκ1 comprising the DPK9 germline gene and the Jκ1 segment for Vκ. The libraries comprise 1 x 1010 individual clones. A subset of the VH and Vκ libraries has been preselected for protein A and generic ligand binding protein L respectively so that most of the clones in the unselected libraries were functional. The sizes of the libraries shown above correspond to the sizes after

40 preselection.

Two or three rounds of selection in mouse, porcine and human serum albumin were performed using subsets of the libraries of the VH and Vκ libraries separately. For each selection, the antigen (i) was coated in an immunotube (nunc) in 4 ml of PBS at a concentration of 100 µg / ml, or (ii) was subjected to biotinylation and then soluble selection was used followed by capture in streptavidin pearls or neutravidin pearls. In

In each case, after the second or third round of selection, the selection DNA was cloned into an expression vector for the production of soluble dAb, and individual colonies were taken. Soluble dAb fragments were produced as described for scFv fragments by Harrison et al. (Methods Enzymol. 1996; 267: 83-109) and for each selection, 96 soluble clones were tested for binding to a range of serum albumins.

Screening of clones for serum albumin binding from a range of species was performed using a

50 Biacore surface plasmon resonance instrument (Biacore AB). A CM-5 Biacore chip was coated with serum albumin of different species at high density in each of 2 to 4 flow cells. The dAb that showed binding to one or more serum albumin of interest were sequenced and expressed on a scale of 50 ml, purified on protein L and then screened at a known concentration for binding to a panel of serum albumin on a CM-5 Biacore chip coated with low serum albumin density in flow cells 2

55 to 4. Several dAb were found that bind to serum albumin from a range of different species, where the preferred candidates are collected, with their binding profiles, in Table 7.

Table 7

HSA (affinity if measured)

RSA (affinity if measured) MSA (affinity if measured) Cyno (affinity if measured)

DOM7h-9

It joins 200 nM joins joins joins

DOM7h-10

joins ND ND ND

DOM7h-11

joins joins joins joins

DOM7h-12

joins ND joins joins

DOM7h-13

joins joins joins

DOM7h-14

Binds 38 nM joins Binds 27 nM Joins 123 nM

In this experiment, dAb that bind to HSA and albumin have therefore been isolated from one or more of a range of non-human species. For example, dAb have been found that bind albumin of (i) human being and mouse, (ii) human being and cynomolgus, (iii) human being and rat and (iv) human being, mouse, rat and cyno.

Example 14

Determination of the serum half-life in rat and mono cynomolgus of fusion proteins of the serum albumin epitope dAb / HA or dAb / myc epitope label and determination of the serum half-life.

Cynomolgus serum anti-albumin dAb were expressed with myc or HA tags at the C-terminus in the

10 periplasm of E. coli and were purified using batch absorption in L-agarose protein affinity resin (Affitech, Norway) for Vk dAb and batch absorption in protein A affinity resin for VH dAb, followed by elution with glycine at pH 2.0. In order to determine the serum half-life, a single i.v. injection was given to groups of 3 cynomolgus macaques. at 2.5 mg / kg of DOM7h-9, DOM7h-11 or DOM7h-14. Blood samples were obtained by serial bleeding of a femoral artery or vein over a period of 21 days and

15 prepared serum from each sample. Serum samples were analyzed by sandwich ELISA using goat anti-HA (Abcam, Cambridge RU) or goat anti myc (Abcam, Cambridge RU) coated on an ELISA plate, followed by detection with L-HRP protein. Standard curves of known dAb concentrations were adjusted in the presence of cynomolgus serum at the same concentration as for the experimental samples to ensure comparability with the test samples. Adjustment with exponential double modeling with a model was used

20 of 2 compartments (using Kaleidograph software (Synergy software, PA, USA).) To calculate t1 / 2β, see Table 8.

The anti-rat serum albumin dAb with myc or HD tags at the C-terminus in the periplasm of E. coli were expressed and purified using L-agarose protein affinity resin batch absorption (Affitech, Norway) followed by elution with glycine at pH 2.0. The dAb were then labeled with 3H using the following method: One vial per protein was prepared: 300 µL of NSP was dispensed into the vial and the solvent was removed under a gentle flow of nitrogen at ≤ 30 ° C. The residue was then resuspended in DMSO (100 L). An aliquot of protein solution (2.5 mL) was added to the DMSO solution and the mixture was incubated for 60 minutes at room temperature. Exactly 2.5 ml of the solution was then loaded onto a previously balanced PD10 column (pre-equilibrated with 25 mL of phosphate buffer saline, PBS) and discarded.

30 eluted. Saline serum with phosphate buffer (PBS, 3.5 mL) was added and the eluate was collected. Thus a protein solution labeled at about 2 mg / mL is obtained. The specific activity of the material was determined and conditioned on effective labeling, the solution was used immediately or stored at -20 ° C until needed.

To determine serum half-life, a single i.v. injection was administered to groups of 4 rats. at 2.5 mg / kg of

35 DOM7h-9, DOM7h-11, DOM7h-13 or DOM7h-14. Blood samples were obtained from a caudal vein over a period of 7 days and the plasma was prepared. 3H levels were determined by liquid scintillation counting and the concentration of labeled proteins in each sample was calculated according to the specific known activity of the protein administered at the beginning of the experiment. Double exponential modeling adjustment was used with a 2-compartment model (using Kaleidograph software (Sinergia software, PA,

40 USA).) To calculate t1 / 2β, see Table 8.

Table 8

Agent

Frame t1 / 2β (cyno) t1 / 2β (rat)

DOM7h-9

Vκ 3.8 days 66 hours

DOM7h-11

Vκ 5.2 days 61 hours

DOM7h-13

Vκ not tested 73 hours

DOM7h-14

Vκ 6.8 days 56 hours

DOM7r-3

Vκ 53 hours

DOM7r-16

Vκ 43 hours

DOM7h-9

Vκ 3.8 days 66 hours

DOM7h-11

Vκ 5.2 days 61 hours

DOM7h-13

Vκ not tested 73 hours

DOM7h-14

Vκ 6.8 days 56 hours

DOM7r-3

Vκ 53 hours

DOM7r-16

Vκ 43 hours

The half-life of albumin in rat and cynomolgus monkey is 53 hours (experimentally determined) and 7-8 days (estimated) respectively. In Table 8 it can be seen that the half-life of dAb DOM7r-3, DOM7h-9, DOM7h-11,

5 DOM7h-13 and DOM7h-14 in rat approximates or is substantially the same as the half-life of albumin in rat. Also, it can be seen that the half-life of dAb DOM7h-11 and DOM7h-14 in cynomolgus approximates or is substantially the same as the half-life of albumin in cynomolgus. The dAb DOM7h-14 has a half-life in rat and cynomolgus that is substantially the same as the half-life of albumin in the two species.

Example 15. Epitope mapping

10 The three domains of human serum albumin have been previously expressed in Pichia pastoris (Dockal Carter and Ruker (1999) J. Biol. Chem. 2000 Feb 4; 275 (5): 3042-50. The same domains are expressed using the vector using Pichia pastoris pPICZaA and in which purification was required to achieve matrix homogeneity of SA Mimetic Blue (supplier: Prometic Biosciences) of Figure 21. The identification of the serum albumin domain bound by the dAb was evaluated by one of two possible methods, immunoprecipitation of domain antibodies and

15 Biacore competition. The results are shown later in Figure 22 and Figure 23.

For immunoprecipitation assay, 1 ml of Pichia pastoris supernatant expressing domain I, II or III of HSA was adjusted to pH 7.4, and mixed with 1 µg of dAb, and 10 µl of Protein A or Protein L agarose (for dAb of VH or VK respectively). The mixture was mixed by inversion for 1 hour to allow complex formation, and then the agarose-bound complex was recovered by centrifugation at 13,000xg for 10

20 minutes, the supernatant was decanted and the sedimented material was washed once with PBS, and recovered by centrifugation. The beads were then resuspended in SDS-PAGE loading buffer containing dithiothreitol (DTT), washed at 70 ° C for 10 minutes, and then passed through 4-12% SDS-PAGE NuPAGE gels (supplier: Invitrogen ), and stained with SimplyBlue safe staining.

For competition testing with Biacore, purified dAb were prepared up to 1 M in HBS-EP at pH 7.4, or 1 M

25 in 50 mM citrate phosphate buffer, 150 mM NaCl, pH 5.0, and when necessary, with 7 µM of purified HSA domain. Preparations are made with Biacore at a flow rate of 30 µl / min on a CM5 chip surface coated with 500-1,000 RU of human serum albumin, and a reference cell used to perform the base subtraction.

Table 9 provides a list of the specific dAb for human serum albumin and the domains or human serum albumin domains with which to perform the mapping (as determined by immunoprecipitation and / or Biacore): Table 9

Clone

H / K HSA domain mapped

DOM7h-1

K Domain II

DOM7h-2

K Nd

DOM7h-6

K Nd

DOM7h-7

K Nd

DOM7h-8

K Domain II

DOM7h-9

K Domain II

DOM7h-10

K Nd

DOM7h-11

K Domain II

DOM7h-12

K Domain II

DOM7h-13

K Domain II

DOM7h-14

K Domain II

DOM7h-21

H Nd

DOM7h-22

H Domain I + III

DOM7h-23

H Nd

DOM7h-24

H Nd

DOM7h-25

H Nd

DOM7h-26

H Nd

DOM7h-27

H Domain III

DOM7h-30

H Domain III

DOM7h-31

H Nd

Nd: not determined

In conclusion, most of the dAb bind to the 2nd HSA domain and therefore are not expected to compete with the binding of human serum albumin to FcRn. Two dAb (DOM7h-27 and DOM7h-30) join Domain III.

dAb

United HSA domain RU binding to HSA at 1 pHM pH 7.4 RU binding to HSA at 1 pHM pH 5.0 His at CDR

DOM7h-1

II 600c 150 no

DOM7h-3

NEITHER 0 0

DOM7h-4

NEITHER 0 0

DOM7h-8

II 1000 250 no

DOM7h-9

II 150 0 CDR1

DOM7h-11

II 250 0 CDR3

DOM7h-12

IIa 55 0 no

DOM7h-13

II 300 40 2 in CDR3

DOM7h-14

II twenty 0 no

DOM7h-22

I + IIIb 100c 0 CDR2

DOM7h-27

III fifty 0 no

DOM7h-30

III 320 35 no

Summary of cartography results of AlbudAb ™ HSA binding epitopes (dAb that specifically bind serum albumin) and Biacore data at pH 7.4 and 5.0.

Example 16 Selection of dAb in vitro in the presence of metabolites

5 Albumin molecules accumulate the effects of exposure to other serum compounds during their life of about 19 days. These effects include the binding of numerous molecules that have an albumin affinity that include but is preferably not limited to cysteine and glutathione transported as mixed disulfides, vitamin B6, δ-bilirubin, hemin, thyroxine, long and medium chain fatty acids and glucose transported in ε-amino groups. In addition, metabolites such as acetaldehyde (a product of ethanol metabolism in the liver), metabolites of

10 fatty acids, acyl glucuronide and bilirubin metabolites. In addition, many drugs such as warfarin, halothane, salicylate, benzodiazepines and others (reviewed in Fasano et al. 2005, IUBMB Life)) and also 1-O-gemfibrocil-β-Dglucuronide bind to serum albumin.

Compounds that are bound to serum albumin often bind at certain sites in the albumin molecule, potentially blocking these sites for the binding of other molecules such as AlbuAbs ™ (a 15 dAb that specifically binds to serum albumin). Binding sites for many ligands have been identified, the main and best characterized binding sites are called "Sudlow 1 site" and "Sudlow 2 site". According to this nomenclature, Site 1 is located in sub-domain IIA, and binds to warfarin and other drugs that are generally bulky heterocyclic anionic molecules. Site 2 is located in subdomain IIIA, and binds aromatic carboxylic acids with an extended conformation, with the negative charge towards one end, such as the ligand

20 of stereotypic site 2, ibuprofen. Secondary binding sites for warfarin and ibuprofen have been identified in domains II and I respectively. There are also other sites and sub-binding sites of these, which means that in circulation, serum albumin exists with a complex mixture of bound ligands, with affinities ranging from 1 x 10-2 M to 1 x 10-8 M. For example, oleic acid is bound to up to 7 sites in SA (J Mol Biol. 2001; 314: 955-60).

25 Human serum albumin has been in crystallized complex with fatty acids (Petitpas I, Grune T, Bhattacharya AA, Curry S. Nat. Struct Biol. (1998) 5: 827-35). The binding sites for these molecules are located in hydrophobic grooves around the surface of SA, with an asymmetric distribution, despite the close triple symmetry of the HSA molecule. Later, the use of several recombinant serum albumin fragments has helped to more accurately assign the contribution of domains to the formation of binding sites (by

30 example: Protein Sci (2000) 9: 1455-65; J Biol Chem. (1999) 274: 29303-10). The displacement of ligands bound from SA plays an important role in drug interactions, for example the warfarin half-life is reduced when it is displaced from the SA by ethanol (J Biol Chem. (2000) 275: 38731-8). The affinity of other drugs for SA is modified by the presence of other drugs at other binding sites. For example, the binding of diazepam at site 2 increases the affinity of site 1 for tenoxicam, as a result of changes

35 conformational in the union. This fact significantly affects the pharmacokinetic properties (Fundam Clin Pharmacol. (1989) 3: 267-79).

Thus, for an SA-binding AlbudAb ™ (a dAb that specifically binds to serum albumin), it is desirable to select one that does not modify the serum albumin binding characteristics for drugs bound to SA. In addition, when it has been shown that drug binding modifies the conformation of SA, it is desirable to have an AlbudAb ™ (a dAb that specifically binds to serum albumin) that binds to SA both in the presence and absence of the drug. These approaches mean that it will be possible to identify an AlbudAb ™ (a dAb that specifically binds to serum albumin) so that there are no significant positive or negative interactions with key pharmaceutical products. Therefore, this example describes a selection of phages to identify the dAb that bind to serum albumin in the presence of compounds and metabolites that are likely to be present together

45 to albumin in vivo. Phage selections are made in the presence of one or more of the metabolites or compounds that are known to interact with serum albumin in vivo. These selections identify the AlbudAb ™ (dAb that specifically bind to serum albumin) that will bind to serum albumin in a way that is unlikely to be hindered by the presence of metabolites or other compounds.

The phage libraries described in Example 1 are used as described in Example 1 for selection

50 against albumin of one or more of a range of species that includes the human being, the cynomolgus monkey, the rat and the mouse. The albumin used as an antigen is different from that described in Example 1 in that it will be pre-incubated overnight with a metabolite or compound at a concentration 10-100 times higher than the albumin itself. This can be done with a single compound or metabolite, or with more than one compound or metabolite. In particular, it may be in the presence of compounds that occupy the albumin sites I or II or both. This metabolite concentration is also present in the buffer used to coat the immunotubes with antigen and in the buffers used during the key stages of the selection. The stages in which the metabolites are present include the MPBS blocking buffer used to block the antigen-coated immunotubes or the biotinylated antigen (for solution selections) and also the buffer in which the phage library is blocked. In this way, when the blocked phages are added to the immunotube or biotinylated antigen, the metabolite concentration is maintained. Therefore, throughout the selection phases in which the phage that binds to albumin is selected, metabolites that can block certain sites in the albumin molecule in vivo are also present, competing with the phage for binding and skewing the selection in favor of those dAb that bind to albumin sites other than those blocked by metabolites.

In another set of selections, alternate batches of selection are performed against serum albumin in the presence and absence of bound compounds or metabolites. This ensures that the dAb capable of binding to serum albumin in the presence and absence of bound compounds are selected. In the two selection schemes, it is possible to select the dAb that are capable of displacing the drug bound to serum albumin, and this is screened by measuring the capacity of AlbudAb ™ (a dAb that specifically binds to serum albumin) of displacing the drug bound to SA. These assays are well established for small molecule drugs, and are easily adapted for this purpose. A variety of methods well known in the art can be used to determine the ability of an AlbudAb ™ (a dAb that specifically binds to serum albumin) to displace drugs bound to SA. They include from equilibrium dialysis, chromatographic methods in serum albumin or immobilized ligands, through NMR analysis. The following example describes the use of the simplest equilibrium dialysis method. The other technically more complex methods will provide essentially the same information.

A serum albumin solution is prepared at a defined concentration in a physiological buffer, for example, 20 mM sodium phosphate buffer, 150 mM NaCl, pH 7.4. The drug is prepared in a similar buffer, and has been synthesized so that it retains its original pharmacological properties, but is radiolabeled, for example with tritium or 14C. The serum albumin binding antibody fragment is prepared at a defined concentration in a similar buffer.

The serum albumin solution is placed in a series of tubes, and an increasing amount of AlbudAb ™ (a dAb that specifically binds to serum albumin) is added, so that the concentration of serum albumin in each tube is fixed (for example at 1% w / v, approximately 150 M), while the concentration of (dAb that specifically binds to serum albumin) ™ is between 0 and 150 M in the tube series. This comprises an experimental set.

A tube or dialysis vessel containing a fixed concentration of the radiolabeled ligand for each set is added to the tube. A concentration between 0.2 and 10 mM may be suitable, depending on the ligand used, its affinity and its solubility.

The membrane cut-off size used for dialysis should be such that serum albumin and AlbudAb ™ (a dAb that specifically binds to serum albumin) do not diffuse through it, but the radiolabeled ligand can diffuse freely. A cut size of 3.5 kDa is sufficient for this purpose.

The mixture is stirred at a fixed temperature, for example 37 ° C, for a fixed period of time, to allow the equilibrium of the radiolabeled drug between the two compartments, for example, 5 hours. After this time, the balance that influences the ability of AlbudAb ™ (a dAb that specifically binds to serum albumin) to bind to serum albumin to inhibit drug binding should be achieved.

The two compartments are samples, and the radioactivity has a scintillation counter. The concentration of albumin bound ligand can be determined by the difference in counts between the two compartments. The stoichiometric binding constant K ’can be calculated from the equilibrium concentration of bound ligand, b, free ligand, c, and albumin, p, according to the equation K’ = b / c (p-b). This presupposes the binding of 1 ligand molecule with a serum albumin molecule.

Binding data can then be measured using a Scatchard plot according to the equation r / c = nk-rk, where r is the albumin fraction to which the ligand is bound (i.e. b / p, and n is the number of binding sites per albumin molecule, and k is the site association constant, and the values of nyk can be determined from the graphs of r / c with respect to r.

When the binding of an AlbudAb ™ (a dAb that specifically binds to serum albumin) blocks the binding of the radiolabeled ligand, this fact will affect the stoichiometric binding constant of the ligand and also the apparent number of binding sites for the ligand. It can be predicted that as AlbudAb ™ (a dAb that specifically binds to serum albumin) will bind at a defined site on the surface of serum albumin, and some ligands have more than one binding site in serum albumin, which not all binding sites will be blocked. In the situation where AlbudAb ™ (a dAb that specifically binds to serum albumin) is specifically bound to a drug that complexes with serum albumin and displaces it, and the drug has a low therapeutic index and is bound to serum, then the cutting affinity to distinguish between a AlbudAb ™ (a dAb that specifically binds to serum albumin) capable of displacing the drug-bound serum albumin from a AlbudAb ™ (a dAb that specifically binds to serum albumin) not capable of displacing serum albumin to the drug would be between 10 nM and 100 nM. This method is illustrated in the following article: Livesey and Lund Biochem

J. 204 (1): 265-272 Binding of branched-chain 2-oxo acids to bovine serum albumin.

Example 17: Generation of a double specific ligand comprising a non-immunoglobulin CTLA-4 serum albumin binding framework by means of CDR grafting

The CDR domains of dAb7h14 are used to construct a non-immunoglobulin framework polypeptide associated with cytotoxic T lymphocytes 4 (CTLA-4) that binds to human serum albumin as follows. The sequences CDR1 (RASQWIGSQLS; SEQ ID NO .:), CDR2 (WRSSLQS; SEQ ID NO.:_) and CDR3 (AQGAALPRT; SEQ ID NO .: _) of dAb7h14 are grafted into a soluble truncated mutant of CTLA-4 that comprises the type V domain of CTLA-4 (as described in WO-99/45110; optionally, a designed form of CTLA4, for example, in which domains A2 and A3 are suppressed) in the substitution of residues of natural CTLA-4 amino acids corresponding to CDR1 (SPGKATE; SEQ ID NO .: _) in the S1-S2 loop (the BC loop), CDR2 (YMMGNELTF; SEQ ID NO .: _) and CDR3 (LMYPPPYYL; SEQ ID NO .: _) in the S5-S6 loop, respectively (for details on the composition and / or structure of the CTLA-4 frame, see WO00 / 60070; WO-99/45110; Metzler et al. Nat Struct. Biol. 4: 527-53; and Nuttall et al. Proteins Struct. Funct. Genet. 36: 217-27, all incorporated herein by reference in their entirety). The expression of this polypeptide derived from CLTA-4 is carried out in an expression system based on pGC, on pPOW or other expression system recognized by the technique, with the expected production of predominantly monomeric soluble proteins. The solubility of the proteins of this polypeptide derived from CTLA-4 is analyzed, and is expected to be superior to the natural extracellular CTLA-4 polypeptide. The ELISA analysis is used to examine whether the purified monomeric polypeptide specifically binds to human serum albumin compared to non-specific antigens and compared to extracellular CTLA-4 derived polypeptides grafted with non-specific polypeptides (eg, structure-substituted somatostatin loop CDR1). Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to an immobilized CTLA-4 derived polypeptide comprising the human serum antialbumin CDR domains of dAb7h14. (One skilled in the art will recognize that binding affinity can be assessed using any appropriate method, including, for example, precipitation of labeled human serum albumin, competitive Biacore assay, etc.). Optionally, the expression of the human serum antialbumin polypeptide CTLA-4 is enhanced by adjusting the coding sequence using overlay and splicing PCR to incorporate preferred codons for E. coli expression. If affinity for low or no human serum albumin is detected (for example, Kd values in the range of M or higher), at least one of several strategies is used to improve the human serum albumin binding properties of the CTLA polypeptide -4 grafted with CDR, which includes any of the following methods that contribute to binding affinity.

Human serum albumin binding of CDR-grafted CTLA-4 polypeptide or polypeptides having dRab7h14 CDR is optimized by mutagenesis, optionally in combination with parallel and / or iterative selection methods as described below and / or as known in the art by other means. The domains of CTLA-4 framework polypeptides surrounding the grafted CDR sequences of dAb7h14 are subjected to randomized and / or NNK mutagenesis, performed as described below. This mutagenesis is performed in the CTLA-4 polypeptide sequence in non-CDR amino acid residues, in order to create new or improved human serum albumin binding polypeptides. Optionally, dAb7h14 CDR polypeptide domains presented in CDR grafted CTLA-4 polypeptide are subjected to mutagenesis, for example, by random mutagenesis, NNK mutagenesis, review mutagenesis and / or other method recognized in the art. PCR is optionally used to perform said mutagenesis methods, with the result of the generation of sequence diversity through sequences directed at CDLA-4 grafted CTLA-4 polypeptides. These approaches are similar to those described below for generating dAb libraries. In addition to random and / or review mutagenesis methods, targeted mutagenesis of target amino acid residues is used in which the structural information establishes specific amino acid residues critical for human serum albumin binding.

CTLA-4 polypeptides comprising grafted dAb7h14 CDR sequences designed as described above are subjected to parallel and / or iterative selection methods to identify those CTLA-4 polypeptides that are optimized for human serum albumin binding. For example, after the production of a library of CTLA-4 polypeptide sequences grafted with CDR dAb7h14, this library of said polypeptides is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for the selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized CTLA-4 derived polypeptides ideally achieving values of human serum albumin binding affinity of Kd in the range of nM or better.

After identification of CTLA-4 derived polypeptides that bind human serum albumin, said polypeptides are then used to generate double specific ligand compositions by any of the methods described below.

Example 18: Generation of a double specific ligand comprising a non-immunoglobulin framework of serum albumin binding CTLA-4 by means of selecting serum albumin binding moieties

A soluble truncated CTLA-4 mutant is presented in a library comprising the type V domain of natural CTLA-4 (as described in WO-99/45110; optionally, a designed form of CTLA-4, for example , in which domains A2 and A3 have been deleted) and which has been designed to contain a region or regions of variability, and is subjected to selection and, optionally, affinity maturation techniques in order to produce shell molecules of non-human serum albumin binding CTLA-4 immunoglobulins for use in the ligands of the invention.

The expression of this polypeptide derived from CLTA-4 is performed in a recognized expression system based on pGC, pPOW or other technique. The protein solubility of this CTLA-4 derived polypeptide is analyzed, and mutagenesis is performed to enhance the solubility of the CTLA-4 derived polypeptide (s) with respect to those of a natural extracellular CTLA-4 polypeptide. ELISA analysis is used to examine whether the purified monomeric polypeptide optionally binds specifically to human serum albumin compared to non-specific individual variable domains comprising a framework derived from CTLA-4, and compared to extracellular polypeptides derived from CTLA- 4 grafted with non-specific polypeptides (for example, CTLA-4 polypeptide with substituted somatostatin in the CDR1 loop structure). Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to an immobilized CTLA-4 derived polypeptide. Optionally, expression of the human serum anti-albumin polypeptide of CTLA-4 is enhanced by adjusting the coding sequence using overlay and splicing PCR to incorporate preferred codons for E. coli expression. After the detection of null or low binding affinity (for example, Kd values in the range of M or higher) of a CTLA-4 polypeptide for human serum albumin, at least one of a set of strategies is employed for imparting human serum albumin binding properties to the CTLA-4 polypeptide, which includes one or more of the following methods that contribute to binding affinity.

Human serum albumin binding of CTLA-4 framework polypeptide or polypeptides is achieved and optimized by mutagenic methods, optionally in combination with parallel and / or iterative selection methods as described below and / or as known by other means in the art. The CTLA-4 polypeptide domains are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed on the entire CTLA-4 polypeptide or on specific sequences in the CTLA-4 polypeptide, optionally targeting amino acids corresponding to CDR (for example, the CDR1 and / or CDR3 regions are randomized, and the resulting polypeptides they are subjected to selection, eg, as described in Example 6 of WO-99/45110). Optionally, specific amino acid residues for which it has been determined or predicted to be structurally important for the presentation of CDR type loop are directed to mutagenesis. Mutagenesis, especially randomized mutagenesis, is performed in order to obtain new or improved human serum albumin binding polypeptides by evolution. PCR is optionally used to carry out said mutagenesis methods, which leads to the generation of a variety of sequences through the sequences directed at the CTLA-4 polypeptides. (These approaches are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is employed in which the structural information establishes specific amino acid residues of CTLA-4 polypeptides that are critical for human serum albumin binding.

CTLA-4 polypeptides designed as described above are subjected to parallel and / or iterative selection methods to identify those CTLA-4 polypeptides that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized CTLA-4 polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for the selection of serum albumin binding dAb from the dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized CTLA-4 derived polypeptides that ideally achieve values of Kd of serum albumin binding affinity in the range of nM or better.

After identification of the CTLA-4 polypeptides that bind human serum albumin, said polypeptides are used to generate double specific ligand compositions by any of the methods described below.

CTLA-4 TYPE V DOMAINS

CTLA-4 is an example of a non-immunoglobulin ligand that is bound to a specific binding partner and also comprises type V domains. These type V domains are distinguished from those of T-cell antibodies or receptors because they have no propensity to join in molecules of type Fv. Said non-immunoglobulin ligand provides an alternative framework for the development of new binding moieties with high affinities for target molecules. Thus, single domain type V binding molecules obtained from CTLA-4 that are soluble are desirable.

The antigen associated with cytotoxic T lymphocytes 4 (CTLA-4) is involved in the regulation of T lymphocytes during the autoimmune response. CTLA-4 is a 44 kDa homodimer expressed primarily and transiently on the surface of activated T lymphocytes, in which it interacts with CD80 and CD86 surface antigens in antigen presenting cells to effect autoimmune response regulation (Waterhouse et al. 1996 Immunol. Rev 153: 183-207, van der Merwe et al. 1997 J Exp Med 185: 393-403). Each monomeric subunit of CTLA-4 consists of an extracellular domain at the N end, transmembrane region and intracellular domain at the C end. The extracellular domain comprises a type V domain at the N end (VLD; approximately 14 kDa molecular weight predicted by homology with the immunoglobulin superfamily) and a stem of approximately 10 residues that connect the VLD with the transmembrane region. The VLD comprises surface loops corresponding to CDR-1, CDR-2 and CDR-3 of an antibody V domain (Metzler et al. 1997 Nat Struct Biol 4: 527-531). Recent structural and mutational studies on CTLA-4 indicate that binding to CD80 and CD86 takes place through the surface of the VLD formed from type V beta chains A'GFCC 'V and also from the MYPPPY sequence highly conserved in the loop of The CDR3 type surface (Peach et al. 1994 J Exp Med 180: 2049-2058; Morton et al. 1996 J. Immunol. 156: 1047-1054; Metzler et al. 1997 Nat Struct Biol 4: 527-531). The dimerization between CTLA-4 monomers takes place through a disulfide bond between cysteine residues (CYS120) in the two stems, which causes the binding of the two extracellular domains, but without any apparent direct association between type domains. V (Metzler et al. 1997 Nat Struct Biol 4: 527-531).

It has previously been shown that the replacement of CDR loop structures in VLDs leads to the production of correctly folded monomeric molecules with altered binding specificities and improved solubility. Accordingly, in some embodiments, a binding moiety is generated comprising at least one monomer type V (VLD) domain derived from CTLA-4, wherein the at least one monomer type V domain is characterized in that at least one structure The CDR loop or part thereof is modified or substituted so that the solubility of the modified VLD improves when compared to the unmodified VLD.

In some embodiments, at least one CDR loop structure or part thereof is modified or replaced so that (i) the size of the CDR loop structure is increased when compared to the corresponding CDR loop structure in the unmodified VLD; and / or (ii) the modification or substitution leads to the formation of a disulfide bond in or between one or more of the CDR loop structures.

A binding moiety comprising at least one monomeric V type domain (VLD) derived from CTLA-4 is described herein, with the at least one monomeric V type domain characterized in that at least one CDR loop structure or part thereof is modified or replaced so that (i) the size of the CDR loop structure is modified when compared to the corresponding CDR loop structure in the unmodified VLD; and / or (ii) the modification or substitution leads to the formation of a disulfide bond in or between one or more of the CDR loop structures.

The size of the CDR loop structure can be increased by at least two, more preferably at least three, more preferably at least six and more preferably at least nine amino acid residues. The modified binding moiety may also show an altered binding affinity or specificity when compared to the unmodified binding moiety. Preferably, the effect of substituting or modifying the CDR loop structure is to reduce or eliminate the affinity of the VLD in one or more natural ligands of the unmodified VLD. Preferably, the effect of substituting or modifying the CDR loop structure is also to change the binding specificity of the VLD (eg, to produce a composition that binds to human serum albumin). Thus, it is preferred that the modified VLD be linked to a specific binding partner (eg, human serum albumin) that is different from that of the unmodified VLD.

The phrase "VLD" is intended to refer to a domain that has structural characteristics similar to the variable heavy chain (VH) or variable light chain (VL) antibody. These similar structural features include CDR loop structures.

As used herein, the term "CDR loop structures" refers to surface regions or loop structures of surface polypeptides as the regions of determining complementarity in V domains of antibodies.

It will be noted that the binding moieties derived from CTLA-4 can be coupled together, chemically or genetically, to form multifunctional or multivalent reagents. For example, the addition of tails at the C-terminus, as in natural CTLA-4 with Cys 20, will produce a dimer. The binding moieties can also be coupled to other molecules for various formulations, including those comprising double specific ligands. For example, CTLA-4 VLDs can comprise a tail of polypeptides at the C-terminus or can be coupled with streptavidin or biotin. CTLA-4 VLDs can also be coupled with radioisotopes, dye markers or other reagents of imaging techniques for the detection and / or in vivo localization of cancers, blood clots, etc. CTLA-4 VLDs can also be immobilized by coupling in insoluble devices and platforms for diagnostic or biosensing applications.

Type V domain of extracellular CTLA-4 can be used. One or more surface loops of the type V domain of CTLA-4 and preferably the CDR1, CDR2 or CDR3 loop structures are replaced by a polypeptide having a binding affinity for serum albumin (eg, CDR dAb7h14 domains and sequences derived therefrom, as illustrated below). It will be noted that these CTLA-4 VLDs can be polyspecific, having affinities directed by their natural surfaces and their modified polypeptide bonds.

One or more of the CDR loop structures of the CTLA-4 VLD may be substituted by one or more CDR loop structures derived from an antibody. The antibody can be obtained from any species. The antibody can be obtained from human, rat, mouse, camel, llama or shark. The CDR1 and CDR3 loop structures can adopt non-canonical conformations that are extremely heterologous in length. The type V domain may also possess a disulfide bond that interconnects the CDR1 and CDR3 loop structures (as found in some camel VHH antibodies) or the CDR2 and CDR3 loop structures (as found in some VHH antibodies of flame).

For in vivo applications it is preferable that the VLDs are homologous to the subject of treatment or diagnosis and that possible xenoantigens are eliminated. Accordingly, it is preferred that VLD molecules for use in clinical applications be substantially homologous to members of the naturally occurring human immunoglobulin superfamily.

Serum albumin binding of CTLA-4 polypeptides (eg, VLDs derived from CTLA-4) can be optimized by selecting a binding moiety with a serum albumin affinity, for example, comprising screening a library of polynucleotides for the expression of a binding moiety with an affinity for serum albumin, where the polynucleotides have undergone mutagenesis leading to a modification or substitution in at least one CDR loop structure in at least one VLD and where the Solubility of the isolated modified VLD improves when compared to the isolated unmodified VLD.

Those skilled in the art will note that within the context of said affinity screening method, any method of randomized or directed mutagenesis can be used to introduce modifications to type V domains. Mutagenesis can be directed mutagenesis. Optionally, the directed mutagenesis involves the substitution of at least one sequence within at least one CDR loop structure using, for example, splicing or other PCR technology.

Those skilled in the art will also note that the polynucleotide library may contain sequences encoding VLDs comprising CDR loop structures that are substantially identical to CDR loop structures present in naturally occurring immunoglobulins and / or sequences encoding VLD which comprise CDR loop structures of unnatural occurrence. Optionally, the screening process involves the presentation of modified type V domains such as fusions of gene III proteins on the surface of bacteriophage particles.

The library may comprise bacteriophage vectors such as pHFA, fd-tet-dog or pFAB.5c containing the polynucleotides encoding type V domains. The screening process may also involve the presentation of modified type V domains in a system. of selection of ribosomal presentation.

Serum albumin binding molecules derived from CTLA-4 provide the following advantages (i) the use of a natural human protein eliminates the need for subsequent humanization of the recombinant molecule, a stage that is often required for protection against the immune system response if used in human treatments; (ii) the domain is naturally monomeric as described above (the incorporation of the Cys120 residue into a tail at the C-terminus leads to the production of a dimeric molecule); and (iii) structural modifications have produced improved expression levels of E. coli.

The initial determination of the structure of natural CTLA-4 allowed modeling and prediction of regions corresponding to regions of CDR antibodies 1, 2 and 3. It was proposed that these areas could be sensitive to mutation or substitution without an effect substantial in the molecular framework and that would therefore allow the expression of a correctly folded molecule. The published CTLA-4 structure (Metzler et al. 1997 Nat Struct Biol 4: 527531) demonstrated the accuracy of these predictions, despite the unexpected separation of CDR1 from the ligand-binding site, and the extensive flexion of CDR3 for form a flat surface contiguous with the ligand binding face.

Type V domains provide a basic framework for the construction of soluble individual domain molecules, in which the binding specificity of the molecule can be designed by modifying the CDR loop structures. The basic framework residues of the type V domain can be modified according to structural characteristics present in camelid antibodies. Camel heavy chain immunoglobulins differ from "conventional" antibody structures in that they consist of VHH chains, (Hamers-Casterman et al. 1993 Nature 363: 446-448). Camelid antibodies consist of two heavy chains, each of which comprises a VHH domain. Several special characteristics allow these antibodies to overcome the double problem of solubility and the inability to present a sufficiently large antigen binding surface.

First, several unconventional substitutions (predominantly hydrophobic or polar in nature) in exposed framework residues reduce the hydrophobic surface, while maintaining the internal beta sheet framework structure (Desmyter et al. 1996 Nat Struct Biol 3: 803-811). In addition, in the three CDR loops several structural features compensate for the loss of antigen-binding surface normally provided by the VL domain. While the CDR2 loop does not differ widely from other VH domains, the CDR1 and CDR3 loops adopt non-canonical conformations that are extremely heterologous in length. For example, the H1 loop can contain anywhere between 2-8 residues compared to the usual five in Ig molecules. However, it is the CDR3 loop that shows the maximum variation: in 17 sequences of referred camel antibodies, the length of this region varies between 7 and 21 residues (Muyldermans et al. 1994 Protein Eng 7: 1129-1135). Third, many VHH domains of camelids have a disulfide bond that interconnects CDR1 and CDR3 in the case of camels and interconnects CDR1 and CDR2 in the case of llamas (Vu et al. 1997 Molec. Immunol. 34: 1121-113 ). The function of this structural characteristic seems to be to maintain the stability of the loop and to provide a more contoured loop conformation, different from the flat one, that allows the attachment to pockets within the antigen and that provides an increased surface area. However, not all camelid antibodies possess this disulfide bond, which indicates that it is not an absolute structural requirement.

A method for generating and selecting individual VLD molecules with new binding affinities for target molecules (eg, human serum albumin) is also described. This method involves the application of well-known molecular evolution techniques for polypeptides derived from CTLA-4. The method may involve the production of phage or ribosomal presentation libraries for the screening of large numbers of mutated polypeptide derivatives of CTLA-4.

The filamentous f bacteriophage genomes are designed so that the phages have, on their surface, proteins such as Ig-like proteins (scFv, Fabs) that are encoded by the DNA that is contained within the phage (Smith, 1985 Science 228: 1315-1317; Huse et al. 1989 Science 246: 1275-81; McCafferty et al., 1990 Nature 348: 552-4; Hoogenboom et al., 1991 Nucleid Acid Res. 19: 4133-4137). Protein molecules may be present on the surface of bacteriophage Fd, covalently coupled with phage coating proteins encoded by gene III, or less frequently with gene VIII. The insertion of antibody genes into the gene III coat protein provides the expression of 3-5 recombinant protein molecules per phage, located at the ends. In contrast, the insertion of antibody genes into gene VIII has the possibility of showing approximately 2,000 copies of the recombinant protein per phage particle, although this is a multivalent system that could mask the affinity of an individual protein presented. Fd phagemid vectors are also used, since they can be easily changed from the presentation of functional fragments of type Ig on the surface of bacteriophage fd to secrete soluble fragments of type Ig in

E. coli. Recombinant protein fusions presented by phage with the N-terminus of the gene III coating protein are made possible by an amber codon strategically placed between the two protein genes. In the amber codon suppressor strains of E. coli, the resulting Ig III gene fusions are anchored in the phage coating.

A selection process based on protein affinity can be applied to any high affinity binding reagents such as antibodies, antigens, receptors and ligands (see, for example, Winter and Milstein, 1991 Nature 349: 293-299, the complete content of which is incorporated herein by reference). Thus, the selection of the maximum affinity binding protein presented in the bacteriophage is coupled with the recovery of the gene encoding that protein. A phage presenting Ig or non-Ig framework can be selected by affinity by binding to similar binding participants covalently coupled with beads or adsorbed on plastic surfaces in a manner similar to ELISA or solid-phase radioimmunoassays. Although almost any plastic surface will adsorb protein antigens, some commercial products have been formulated especially for this purpose, such as Nunc immunotubes.

Ribosomal presentation libraries involve de novo synthesized polypeptides in cell-free translation systems and presented on the surface of ribosomes for selection purposes (Hanes and Pluckthun, 1997 Proc. Natl. Acad. Sci. USA 94: 4937-4942; He and Taussig, 1997 Nucl. Acids Res. 25: 5132-5134). The "cellless translation system" comprises ribosomes, soluble enzymes required for protein synthesis (usually of the same cell as ribosomes), transfer RNA, adenosine triphosphate, guanosine triphosphate, a ribonucleoside triphosphate regenerator system (such as phosphoenol pyruvate and pyruvate kinase) and the salts and buffer required to synthesize a protein encoded by an exogenous mRNA. The translation of polypeptides can be performed under conditions that keep the polysomes intact, that is, in which the ribosomes, the mRNA molecule and the translated polypeptides are associated in a single complex. This leads in practice to the "ribosome presentation" of the translated polypeptide. For selection, the translated polypeptides, in association with the corresponding ribosome complex, are mixed with a target molecule (e.g., serum albumin) that is bound to a matrix (e.g., Dynabeads). The ribosomes presenting the translated polypeptides will bind to the target molecule and these complexes can be selected and the mRNA reamplified using RT-PCR.

Although there are several alternative strategies for modifying binding molecules, the general strategy for all presented proteins is guided by a pattern in which individual binding reagents are selected from presentation libraries for affinity to their similar ligand and / or receptor. The genes encoding these reagents are modified by any one or a combination of a series of in vivo and in vitro mutation strategies and is constructed as a new gene pool for the presentation and selection of the affinity binding molecules.

LIPOCALINS

Example 19: Generation of a double specific ligand comprising a non-immunoglobulin framework of serum albumin binding lipocalins by means of selecting serum albumin binding moieties

Biline-binding protein (BBP), a lipocalin obtained from Pieris brassicae can be remodeled by combinatorial protein design so that it recognizes a human serum albumin. To this end, the natural BBP is subjected to library selection and, optionally, affinity maturation in order to produce human serum albumin-binding BBP molecules for use in the specific double ligands described herein.

The ability of a natural BBP to bind to human serum albumin is initially determined by means of a Biacore assay, as described below for CTLA-4 derived polypeptides. (One skilled in the art will recognize that binding affinity can be evaluated using any appropriate method, including, for example, precipitation of labeled human serum albumin, competitive Biacore assay, etc.). After detection of zero or low binding affinity (for example, Kd values in the range of M or greater) of BBP for human serum albumin, at least one of a set of strategies is used to impart binding properties to human serum albumin to BBP, which includes one or more of the following methods that contribute to binding affinity.

Binding to human serum albumin of BBP and polypeptide or polypeptides derived from BBP is achieved and optimized by mutagenic methods, optionally in combination with parallel and / or iterative selection methods as described below and / or as known by others. means in the art. The BBP polypeptide domains are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed primarily on the BBP polypeptide (or derived from BBP) and / or is performed on specific sequences within the BBP polypeptide, which includes 16 amino acid residues identified as residing in the center of the natural BPP binding site, which consists of four loops at the top of an eight-chain beta barrel (Beste et al. 1999 Proc. Natl Acad. Sci. USA 96: 1898-903). Optionally, said mutagenesis procedures are randomized in order to obtain new or improved polypeptides binding to human serum albumin; and multiple batches of mutagenesis can be performed during the process of creating a BBP that is optimally bound to human serum albumin. PCR is optionally used to perform mutagenesis methods, which results in the generation of a variety of sequences through sequences directed within the BBP polypeptides (or BBP derivatives). (These strategies are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is employed in which the structural information establishes specific amino acid residues of BBP polypeptides (or BBP derivatives) that will be critical for human serum albumin binding.

BBP polypeptides (or BBP derivatives) designed as described above are subjected to parallel and / or iterative selection methods to identify those BBP polypeptides that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized BBP polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below. for the selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made with respect to serum albumin immobilized in immunotubes or biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized BPP derived polypeptides that ideally reach values of Kd of serum albumin binding affinity in the range of nM or better.

After identification of the BBP polypeptides that bind human serum albumin, said polypeptides are then used to generate double specific ligand compositions by any of the methods described below.

Lipocaline framework proteins

Lipocalins (Pervaiz and Brew, FASEB J. 1 (1987), 209-214) form a family of small, often monomeric, secretory proteins that have been isolated from various organisms, and whose physiological function consists in the storage or transport of different ligands as well as more complex biological functions (Flower, Biochem. J. 318 (1996), 1-14). Lipocalins show a relatively low mutual sequence similarity and their belonging to the same structural family of proteins was first determined by X-ray structure analysis (Sawyer et al., Nature 327 (1987), 659).

The first known spatial structure lipocalin was the retinol-binding protein, Rbp, which carries the transport of water-soluble vitamin A in the blood serum (Newcomer et al., EMBO J. 3 (1984), 1451-1454). Shortly after, the tertiary structure of the bilin-binding protein, Bbp, was determined from the Pieris brassicae butterfly (Huber et al., J. Mol. Biol. 195 (1987), 423-434). The essential structural characteristics of this class of proteins are illustrated in the spatial structure of this lipocalin. The central element in the folding architecture of lipocalins is a sheet structure folded in cylindrical β, called β barrel, which is formed by up to eight β-chain antiparallels arranged almost circularly.

This super-secondary structural element can also be seen as a "sandwich" arrangement of two four-chain β-sheet structures. The additional structural elements are an extended segment at the amino end of the polypeptide chain and an α helix near the carboxy end, which follows from an extended segment. However, these additional features are not necessarily revealed in all lipocalins. For example, an important part of the segment at the N-terminus is omitted in the epididymal retinoic acid binding protein (Newcomer, Structure (1993) 1: 7-18). Additional peculiar structural elements are also known, such as, for example, membrane anchors (Bishop and Weiner, Trends Biochem. Sci. (1996) 21: 127) that are present only in certain lipocalins.

The β barrel is closed at one end by dense packing of amino acids as well as by loop segments. At the other end, the β barrel forms a binding bag in which the respective lipocalin ligand forms a complex. There are eight neighboring antiparallel β chains connected in the form of respective pairs by fork-type flexures in the polypeptide chain which, together with the adjacent amino acids, are still partially located in the region of the cylindrical β-folded sheet structure, each of the which forms a loop element. The binding bag for ligands is formed in total by these four peptide bonds. In the case of Bbp, biliverdin IXγ complex form in this binding bag. Another typical ligand for lipocalins is vitamin A in the case of Rbp as well as β-lactoglobulin (Papiz et al., Nature 324 (1986), 383-385).

As described, for example, in the US publication. No. 20060058510, members of the lipocalin polypeptide family can be used to produce a class of molecules called "anticalins" designed to recognize new ligands by mutating amino acids that are located in the region of the four peptide bonds at the end of the structure. in cylindrical folded sheet β, and which are characterized in that they bind to given ligands (eg, human serum albumin) with an affinity that can be determined.

The ligand binding sites of lipocalins have a simpler construction than those of immunoglobulins. The lipocalin polypeptides comprise only one ring of 8 antiparallel β chains: the β barrel. This cyclic β-folded sheet structure is preserved in the folding of lipocalin proteins. The binding site is formed in the entrance region of the β barrel by means of the four peptide bonds, each of which connects with two β chains adjacent to each other. These peptide bonds can vary significantly in structure between individual members of the lipocalin family.

To use a lipocalin polypeptide as a non-immunoglobulin framework, one or more of the four peptide bonds that form the ligand binding site of a lipocalin is subjected to mutagenesis, followed by the choice, that is, the selection of those variants. of proteins (muteins), which show the desired binding activity for a given ligand. The lipocalin muteins obtained in this way have been called "anticalins".

The four peptide bonds of lipocalins that, during anticalin production, are modified in their sequence by mutagenesis, are characterized by those segments in the BBP linear polypeptide sequence comprising amino acid positions 28 to 45, 58 to 69, 86 to 99 and 114 to 129 of Bbp. Each of these sequence segments begins before the C-terminus of one of the β chains conserved on the open side of the β barrel, includes the actual peptide hairpin and ends after the N-terminus of the analogously conserved β chain which follows in the sequence.

Sequence alignments or structural overlays allow assigning the sequence positions given for Bbp to the other lipocalins. For example, sequence alignments corresponding to the published alignment of Peitsch and Boguski (New Biologist 2 (1990), 197-206) reveal that the four ApoD peptide bonds include amino acid positions 28 to 44, 59 to 70, 85 to 98 and 113 to 127. It is also possible to identify the corresponding peptide bonds in the new lipocalins that are suitable for mutagenesis in the same way.

In some cases, the relatively weak sequence homology of lipocalins may be problematic in determining conserved β chains. Therefore it is crucial that the polypeptide sequence is capable of forming the cyclic β folded sheet structure formed by 8 antiparallel β chains. This can be determined using structural analysis methods such as protein crystallography or multidimensional nuclear magnetic resonance spectroscopy.

In non-Bbp lipocalins, such as ApoD or Rbp, the sequence segments suitable for mutagenesis can be easily longer or shorter than those of Bbp based on the individually variable structure of the peptide bonds. It may even be advantageous to further modify the length of sequence segments by deletion or insertion of one or more amino acids. In some embodiments, these amino acid positions corresponding to sequence positions 34 to 37, 58, 60, 69, 88, 90, 93, 95, 97, 114, 116, 125, and 127 of Bbp are mutated. Correspondingly, in the case of ApoD, sequence positions 34 to 37, 59, 61, 70, 87, 89, 92, 94, 96, 113, 115, 123 and 125 are preferred for mutagenesis. However, for anticalin production, not all sequence positions listed above must undergo mutagenesis.

Other lipocalins are also suitable as an underlying structure for the production of anticalins. Preferably, the Rbp, Bbp or ApoD lipocalins are used, which at present have already been thoroughly studied biochemically. The use of lipocalins of human origin is especially preferred for the production of anticalins. This applies especially when an application of the resulting anticalin or anticalins is intended for humans since, for example, in diagnostic or therapeutic applications in vivo, a minimal immunogenic effect should be expected compared to the lipocalins of other organisms. However, other lipocalins as well as lipocalins that are possibly pending discovery may prove to be especially advantageous for the production of anticalins. Artificial proteins can also be used with a folding element that is structurally equivalent to the β barrel of lipocalins.

Preferably the anticalin molecules described herein should be able to bind to the desired ligand (e.g., human serum albumin) with an affinity that can be determined, that is, with a constant affinity of at least 105 M-1. Affinities below this value in general can no longer be measured exactly with common methods and are therefore of secondary importance for practical applications. They prefer

M-1

especially the anticalinas that bind to the desired ligand with an affinity of at least 106, which corresponds to a dissociation constant for the complex of 1 M. The binding affinity of an anticalin to the desired ligand can be measured by the person skilled in the art by a multitude of methods, for example by fluorescence titration, by competitive ELISA or by surface plasmon resonance technique.

Lipocalin cDNA, which can be produced and cloned by the person skilled in the art by known methods, can serve as a starting point for mutagenesis of the peptide loop, as described for example for Bbp (Schmidt and Skerra, Eur. J. Biochem. 219 (1994), 855-863). Alternatively, genomic DNA can also be used for gene synthesis or a combination of these methods can be applied. For the mutagenesis of the amino acids in the four peptide bonds, the person skilled in the art has at his disposal the various known methods for site-directed mutagenesis or for mutagenesis by means of the polymerase chain reaction. The method of mutagenesis can be characterized, for example, because mixtures of synthetic oligodeoxynucleotides can be used for the introduction of the mutations, which have a degenerate base composition in the desired positions. The implementation of the nucleotide constituent blocks with reduced base pair specificity, such as inosine, is also an option for the introduction of mutations in the sequence segment or the amino acid positions chosen. The procedure for mutagenesis of ligand binding sites is simplified compared to antibodies, since for the lipocalins only four segments of sequences should be manipulated for this purpose, instead of six, which corresponds to the four peptide loops mentioned above.

In the site-directed random mutagenesis methods that implements synthetic oligodeoxynucleotides, the relevant amino acid positions in the structure of the lipocalins that must be mutated can be determined in advance. The ideal selection of the amino acid positions that will be mutated may depend on one part of the lipocalin used, and on the other part of the desired ligand (eg, human serum albumin). It may be useful to maintain the total number of mutated amino acid positions in a single experiment at sufficiently low values so that the collection of variants obtained by mutagenesis, that is, the so-called library, can be in its entirety or, at least in a representative selection of the same, to be carried out as completely as possible in its combinatorial complexity, not only at the level of the coding nucleic acids, but also at the level of the gene products.

It is possible to choose the positions of the amino acids that will be mutated significantly especially when there is structural information related to the lipocalin itself that will be used, such as BBP and Rbp or at least relative to a lipocalin with a similar structure, such as for example in the case of ApoD. The set of amino acid positions chosen may also depend on the characteristics of the desired ligand. It may also be advantageous to exclude the positions of individual amino acids in the region of the ligand binding bag from mutagenesis if, for example, they are shown to be essential for the efficiency of folding or the stability of protein folding. Specific oligonucleotide-based methods of lipocalin mutagenesis are described, for example, in the US publication. No. 20060058510, whose complete content is incorporated herein by reference.

After expressing the nucleic acid coding sequences subjected to mutagenesis, the clones that carry the genetic information for anticalines that bind to a given ligand (e.g., human serum albumin) can be selected from the different library clones obtained. For the selection of these clones, the expression strategies and the selection strategies chosen can be applied. Methods of this type have been described in the context of the production or design of recombinant antibody fragments, such as the "phage display" technique or the "colony screening" methods (Skerra et al., Anal. Biochem. 196 (1991), 151-155).

Descriptions of "phage display" techniques are found, for example, in Hoess, Curr. Opin. Struct. Biol. 3 (1993), 572-579; Wells and Lowman, Curr. Opin. Struct. Biol. 2 (1992), 597-604; and Kay et al., Phage Display of Peptides and Proteins - A Laboratory Manual (1996), Academic Press. Briefly, in an exemplary embodiment, plasmids are produced that effect the expression of the lipocalin mutated structural gene as a fusion protein with a signal sequence at the N-terminus, preferably the OmpA signal sequence, and with the coating protein pIII of phage M13 (Model and Russel, in "The Bacteriophages", Vol. 2 (1988), Plenum Press, New York, 375-456) or fragments of this coating protein, which are incorporated into the phage coating, in the C-terminus. To produce the fusion proteins, the fragment at the C-terminus ApIII of the phage coating protein is preferably used, containing only amino acids 217 to 406 of the natural pIII coating protein. A fragment at end C from pIII in which the cysteine residue at position 201 is absent or substituted by another amino acid is especially preferred. The additional description of the phage display methods, the selection methods, etc., that can be applied to lipocalins in production of "anticalins" that possess specific binding properties are detailed, for example, in US publication. No. 20060058510, whose complete content is incorporated herein by reference.

Anticalins can be identified and produced, for example, using the methods described above, which possess high affinity for a given ligand (eg, human serum albumin). Ligand binding constants of more than 106 M-1 can be achieved for anticalins, even in cases where a new ligand does not behave in any way with biliverdin IXγ, the original Bbp ligand (see US publication No. 20060058510). These affinities for new ligands that can be achieved with anticalins are comparable with the affinities that are known for antibodies to the secondary immune response. In addition, there is the possibility of subjecting the produced anticalins to optionally additional random mutagenesis in order to select even higher affinity variants from the new library thus obtained. The corresponding procedures have already been described for the case of recombinant antibody fragments for "affinity maturation" purposes (Low et al., J. Mol. Biol. 260 (1996), 359-368; Barbas and Burton, Trends Biotechnol 14 (1996), 230-234) and can also be applied to anticalins in a corresponding manner by the person skilled in the art.

Staphylococcal protein A (SPA) / Affibody

Example 20: Generation of double specific ligand comprising a framework of Affibody non-immunoglobulins (staphylococcal protein A (SPA)) binding to serum albumin by selecting serum albumin binding moieties

The Z domain of Staphylococcal Protein A (SPA) is subjected to library selection techniques and, optionally, affinity maturation in order to produce non-immunoglobulin shell molecules derived from human serum albumin binding SPA (called "molecules" affibody ") for use in the specific double ligands described herein.

A real-time binding analysis is performed by Biacore to assess whether human serum albumin is specifically bound to immobilized SPA polypeptide. (One skilled in the art will recognize that binding affinity can be assessed using any appropriate method, including, e.g., precipitation of labeled human serum albumin, competitive Biacore assay, etc.). After detection of zero or low binding affinity (for example, Kd values in the range of M or higher) of an unaltered SPA polypeptide for human serum albumin, at least one of a set of strategies is imparted to impart Human serum albumin binding properties to the SPA polypeptide, which include one or more of the following methods designed to impart and / or improve the binding affinity of the molecule for the target antigen.

Human serum albumin binding of SPA polypeptide or framework polypeptides is achieved and optimized by mutagenic methods, optionally in combination with parallel and / or iterative selection methods as described below and / or known by other means in The technique. The framework polypeptide domains of SPA are subjected to randomized and / or NNK mutagenesis, performed as described below. Such mutagenesis is performed primarily in the Z domain of the SPA polypeptide or on specific sequences within the SPA polypeptide, for example, on 13 surface residues accessible to solvents in the Z domain as identified in Nord et al. (1997 Nat. Biotechnol. 15: 772-77), and optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. Optionally, PCR is used to perform these mutagenesis methods, which results in the generation of a variety of sequences in targeted sequences in the SPA polypeptides. (These strategies are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is employed where the structural information establishes specific amino acid residues of SPA polypeptide that are critical for the binding of human serum albumin. In some embodiments, repertoires of mutant Z domain genes are assembled and inserted into a phagemid vector adapted for monovalent phage display. The libraries they comprise, e.g. For example, millions of transformants are constructed using, e.g. eg, degeneration of NN (G / T) or (C / A / G) alternative NN for mutagenesis.

SPA polypeptides designed as described above are subjected to parallel and / or iterative selection methods to identify said SPA polypeptides that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized SPA polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for the selection of serum albumin binding dAb from dAb libraries. Bioseparation is performed by adsorption against the target human serum albumin protein to achieve significant enrichment of serum albumin binding SPA molecules. The selected clones are subsequently expressed in E. coli and analyzed by SDS-PAGE, circular dichroism spectroscopy and human serum albumin binding studies by biospecific interaction analysis. SPA molecules (affibody molecules) that bind to human serum albumin are expected to have a secondary structure similar to the natural Z domain and have micromolar dissociation constants (Kd) for their respective targets in the range of M or better (p e.g., nM or pM).

Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), where fully optimized SPA derived polypeptides ideally reach Kd values. of serum albumin binding affinity in the range of nM or better.

After identification of SPA polypeptides that bind human serum albumin, said polypeptides are used to generate specific double ligand compositions by any of the methods described below.

Staphylococcal protein Affibody molecule polypeptides A (SPA)

Surfaces exposed to bacterial receptor solvents can be addressed by random mutagenesis followed by phenotypic selection in order to impart, e.g. eg, the binding affinity for serum albumin to said receptor molecules. Such proteins may be abnormally stable, which makes them suitable for various applications (Alexander et al. (1992) Biochemistry 31: 3597-3603). In particular, for bacterial receptors that contain structures in helical groups, the conformation can be expected to be tolerant to changes in the side chains of non-participating residues at the helix packing interfaces. Examples of such molecules are the relatively small (58 residues) B-binding domain to Staphylococcal A protein IgG (SPA) and the synthetic analogue of domain B, denoted as Z domain (Nilsson et al. (1987) Protein Engineering 1: 107 -113).

The Z domain derived from SPA is the main SPA domain used as a framework to construct domain variants with new binding properties (see, e.g., documents WO-00/63243 and WO-95/19374, incorporated into the present memory as a reference in its entirety). The Z domain of SPA is a cysteine-free triple helix group domain of 58 amino acid residues that is used as a framework for the construction of combinatorial phagemid libraries from which variants that target desired molecules are selected (e.g. eg, human serum albumin) using phage display technology (Nilsson et al. 1987 Protein Eng. 1: 107-113; Nord et al. 1997 Nat. Biotechnol. 15: 772-777; Nord et al. 2000 J. Biotechnol. 80: 45-54; Hansson et al. 1999 Immunotechnology 4: 237-252; Eklund et al., 2002 Proteins 48: 454-462; Rönnmark et al. 2002 Eur. J. Biochem. 269: 2647-2655) . Said target binding variants, called "affibody" molecules, are selected as target protein binders by phage display from combinatorial libraries where 13 side chains on the surface of helices 1 and 2 have been randomized (Q9, Q10, N11 , F13, Y14, L17, H18, E24, E25, R27, N28, Q32 and K35) in the Z domain (Lendel et al. 2006 J. Mol. Biol. 359: 1293-304). The simple and robust structure of these affibody molecules, together with their low molecular weight (7 kDa), makes them suitable for a wide variety of applications. Documented efficacy has been demonstrated in bioprocess and laboratory scale bioseparations (Nord et al. 2000 J. Biotechnol. 80: 45-54; Nord et al. 2001 Eur. J. Biochem. 268: 42694277; Gräslund et al. 2002 J. Biotechnol. 99: 41-50), and promising results have been obtained by evaluating affibody molecule ligands as detection reagents (Karlström and Nygren 2001 Anal. Biochem. 295: 22-30; Rönnmark et al. 2002 J. Immunol. Methods 261: 199-211), to design adenoviral tropism (Henning et al. 2002 Hum. Gene Ther. 13: 1427-1439) and to inhibit receptor interactions (Sandström et al. 2003 Protein Eng. 16: 691- 697). Thus, affibody molecule ligands designed that, e.g. eg, human serum albumin binding are desirable components of certain double specific ligand compositions of the present invention.

Polypeptide libraries obtained from the Z domain of staphylococcal protein A can be generated by any method of mutagenesis as is known in the art and / or as described below. After the creation of such polypeptide libraries, variants capable of binding to desired target molecules (e.g., human serum albumin) can be efficiently selected and identified using, for example, in vitro selection technologies as phage display (Dunn 1996 ; Smith and Patrenko 1997; Hoogenboom et al. 1998), ribosomal presentation peptides (Hanes and Pluckthun 1997; He and Taussig 1997) in plasmids (Schatz 1993) or bacterial presentation (Georgiou et al. 1997). For these selections, a correlation between library size (complexity) and the probability of isolating binders of higher affinities (KD = 10-8 M or less) has been theoretically considered (Perelson and Oster 1979) and experimentally demonstrated ( Griffiths et al. 1994; Vaughan et al. 1996; Aujame et al. 1997).

Affibody molecules have several advantages over traditional antibodies, for example (i) a lower manufacturing cost; (ii) smaller size; (iii) increased stability and robustness; and (iv) ability to recombinantly occur in a bacterial host, or by chemical synthesis, which obviates the risk of viral contamination.

An affibody molecule is a polypeptide that is a derivative of a staphylococcal protein A (SPA) domain, said SPA domain being the B or Z domain, where a series of amino acid residues have been replaced by other amino acid residues, being performed said substitution without substantial loss of the basic structure and stability of said SPA domain, and said substitution producing an interaction ability of said polypeptide with at least one target antigen domain (eg, human serum albumin). The number of substituted amino acid residues could be from 1 to about 30, or from 1 to about 13. Other possible ranges are from 4 to about 30; from 4 to about 13; 5 to about 20, or 5 to about 13 amino acid residues. The skilled person will understand, p. eg, according to Nord et al. 1997 Nat. Biotechnol. 15: 772777, that the amino residues preferably located on the surface of the Z domain can be substituted, while the core of the group should be kept constant to preserve the structural properties of the molecule.

A procedure for the manufacture of an affibody molecule, e.g. eg, in WO-00/63243, and for the purposes of the present invention could imply, e.g. For example, the following stages: (i) presentation, p. ex. by phage display (for review, see, e.g., Kay, K. et al. (eds.) Phage Display of Peptides and Proteins: A Laboratory Manual, Academic Press, San Diego, ISBN 80-0), presentation ribosomal (for review, see eg Hanes, J. et al. (1998) Proc. Natl. Acad. Sci. USA 95: 14130-14135) or cell presentation (for review, see eg Daugherty, PS et al. (1998) Protein Eng. 11: 825-832), polypeptide variants of a protein library comprising a repertoire of polypeptide variants derived from the SPA B or Z domain;

(ii) selection of clones expressing polypeptides that bind human serum albumin; and (iii) production by recombinant expression of the selected clones or by chemical synthesis.

Avimer

Example 21: Generation of double specific ligand comprising a serum albumin binding avimer by means of CDR grafting

The CDR domains of dAb7h14 are used to construct an avimeric polypeptide that binds to human serum albumin in the following manner. The sequences CDR1 (RASQWIGSQLS; SEQ ID NO.:__), CDR2 (WRSSLQS; SEQ ID NO.:__), and CDR3 (AQGAALPRT; SEQ ID NO.:__) of dAb7h14 are grafted into a C2 monomer (described in U.S. Patent Publication No. 2005/0221384, incorporated herein by reference in its entirety) in residues 17-28, 49-53 and 78-85, respectively, which constitute the loop regions 1, 2 and 3, respectively of the C2 monomer. Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to immobilized C2-derived monomer polypeptide comprising the human serum anti-albumin CDR domains of dAb7h14. (One skilled in the art will recognize that binding affinity can be assessed using any appropriate method, including, e.g., precipitation of labeled human serum albumin, competitive Biacore assay, etc.). If a low or no human serum albumin affinity is detected (e.g., Kd values in the range of M or higher), at least one of a set of strategies is used to improve serum albumin binding properties C2-grafted human monomer in CDR (and / or avimer dimers, trimers and other higher order iteration compositions), which includes any of the following methods that contribute to binding affinity.

The length or lengths of regions grafted with CDR dAb7h14 of the initial C2 monomeric polypeptide (and / or of dimeric, trimer, etc., produced iteratively) polypeptides corresponding to loop regions exposed to solvents in the native C2 monomer (and / or in other native monomers used in the avimer compositions) are adjusted through the use of linker group polypeptides. For example, the CDR2 peptide sequence of nine amino acid residues of dAb7h14 can be extended to 13 amino acid residues in length using amino acid linker groups of, e.g. eg, zero to four length residues located at one and / or the two flanks at the N or C end of the CDR3 polypeptide sequence dAb7h14, thereby achieving a total grafted peptide sequence length of 13 amino acids in the CDR3 grafted domain corresponding to loop 3 of the C2 monomeric polypeptide. Said use of linker group polypeptide or polypeptides is optionally combined with mutagenesis of linker group sequences, CDR sequences and / or non-CDR C2 monomeric polypeptide sequences (e.g., using mutagenic optimization procedures as described below. ), in order to improve the ability of human serum albumin binding of CD2-grafted polypeptide or C2 monomeric polypeptides (e.g., by optimization of C2 and CDR monomer polypeptide sequences in CDR-grafted C2 monomeric polypeptides) . The polypeptide linker groups used for this purpose have a predetermined sequence, or, optionally, are selected from a population of randomized polypeptide linker group sequences by titration of the human serum albumin binding capabilities of CR2-grafted C2 monomeric polypeptides containing linker group Optimization methods are performed in parallel and / or iteratively. Parallel and iterative optimization processes (eg, affinity maturation) employ selection methods as described below and / or as known in the art as useful for optimizing polypeptide binding properties.

Binding to human serum albumin of CD2-grafted polypeptide or C2 monomeric polypeptides (and / or higher order compositions produced iteratively from dimer, trimer, etc., from avimers, or additional individual monomers contributing thereto) that have CDR dAb7h14 it is optimized by mutagenesis, optionally in combination with parallel and / or iterative selection methods as described below and / or known by other means in the art. For the exemplary C2 monomeric shell polypeptide, the domains surrounding the grafted CDR dAb7h14 polypeptide sequences are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is optionally performed in the polypeptide sequence of the C2 monomer on selected amino acid residues as indicated, e.g. e.g., in U.S. Patent Publication No. 2005/0221384, or is optionally performed on all non-CDR amino acid residues, and optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. Optionally, dAb7h14 CDR polypeptide domains presented in the CDR-grafted C2 monomeric polypeptide are subjected to mutagenesis by, e.g. eg, random mutagenesis, NNK mutagenesis, review mutagenesis and / or other method recognized by the art. PCR is optionally used to perform said mutagenesis methods, which results in the generation of a variety of sequences through sequences directed at the C2-grafted C2-grafted monomeric polypeptides. These strategies are similar to those described below for the generation of dAb libraries. In addition to the methods of random and / or review mutagenesis, directed mutagenesis of targeted amino acid residues is used where the structural information establishes specific amino acid residues that are essential for the binding of human serum albumin.

C2 monomeric polypeptides (and / or iteratively produced avimer compositions comprising individual monomers) comprising grafted CDR dAb7h14 CDR sequences as described above are subjected to parallel and / or iterative selection methods to identify those C2 monomeric polypeptides (and avimer compositions) that are optimized for binding to human serum albumin. For example, after the production of a library of C2 monomeric polypeptide sequences grafted with CDR dAb7h14, this library of said polypeptides is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, such as It is described below for the selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized avimers comprising C2 derived monomers that ideally achieve Kd values of serum albumin binding affinity in the range of nM or better.

Upon identification of polypeptides derived from C2 monomers that bind human serum albumin, the properties of human serum binding of said initial monomers can be further enhanced by combining said monomers with other monomers, followed by mutagenesis and / or additional selection, thus forming a composition of avimers that possess specific affinity for human serum albumin. After the identification of an avimer composition that has an affinity for human serum albumin, said avimer polypeptides are used to generate double specific ligand compositions by any of the methods described below.

Example 22: Generation of double specific ligand comprising a non-immunoglobulin framework of serum albumin binding avimers by means of selecting serum albumin binding moieties

The natural C2 monomeric polypeptide as indicated is subjected to library selection techniques and, optionally, affinity maturation, and then combined with an additional monomer (e.g., a fibronectin monomer, for which the affinity to human serum albumin can optionally be optimized in parallel) and optionally it is subjected iteratively to library selection and, optionally, affinity maturation techniques in order to produce a human serum albumin binding non-immunoglobulin framework molecule for its use in the specific double described herein.

Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to an immobilized C2 monomeric polypeptide (and / or an iteratively produced avimer molecule). After detection of null or low binding affinity (e.g., Kd values in the range of M

or higher) of a C2 monomeric polypeptide for human serum albumin, at least one of a set of strategies is used to impart human serum albumin binding properties to the C2 monomeric polypeptide, which includes one or more of the following methods that contribute to the binding affinity

Human serum albumin binding of C2 monomeric polypeptide or polypeptides (and / or iteratively produced dimer, trimer, etc. molecules) is achieved and optimized by mutagenic methods, optionally in combination with parallel selection methods and / or iterative as described below and / or known by other means in the art. The domains of monomeric C2 polypeptides are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed on the entire C2 monomeric polypeptide or on specific sequences in the C2 monomeric polypeptide on selected amino acid residues as indicated, e.g. e.g., in U.S. Patent Publication No. 2005/0221384, and is optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. PCR is optionally used to perform said mutagenesis methods, which results in the generation of a variety of sequences through sequences directed at the C2 monomeric polypeptide molecules and / or avimers. (These strategies are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is employed where the structural information establishes specific amino acid residues of C2 and / or avimer monomer molecules that are essential for the binding of human serum albumin.

C2 monomeric polypeptides designed as described above are subjected to parallel and / or iterative selection methods to identify those C2 monomeric and / or avimeric polypeptide molecules that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized C2 monomer polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below. for selection of serum albumin binding dAb from dAb libraries. Optionally, the selection batches may include iterations within which additional subunits of monomers are added to form a new avimer molecule. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized avimers comprising C2 derived monomeric polypeptides that achieve ideally Kd values of serum albumin binding affinity in the range of nM or better.

Upon identification of polypeptides derived from C2 monomers that bind human serum albumin, the human serum binding properties of said initial monomers can be further enhanced by combining said monomers with other monomers, followed by additional mutagenesis and / or selection , thus forming a composition of avimers that possess specific affinity for human serum albumin. After identification of an avimer composition that has an affinity for human serum albumin, said avimer polypeptides are used to generate double specific ligand compositions by any of the methods described below.

Production and use of avimer polypeptides

The avimers are evolved from a large family of human extracellular receptor domains by in vitro exon shuffling and phage display, which generates multidomain proteins with binding and / or inhibitory properties. The binding of multiple independent binding domains (selected, e.g., iteratively for binding to a target protein, e.g., human serum albumin) creates avidity and leads to an improvement in affinity and specificity in comparison with conventional single epitope binding proteins. Other potential advantages include simple and efficient production of specific multi-target molecules in E. coli, improved thermostability and protease resistance. Avimers that have affinities lower than nM against a target protein can be produced. For example, an avimer has been produced that inhibits interleukin 6 with IC50 0.8 pM in cell-based assays and has been characterized as biologically active (Silverman et al. 2005 Nature Biotechnology 23: 1556-1561; see also, for example , published U.S. Patent Applications No. 2005/0221384, 2005/0164301, 2005/0053973 and 2005/0089932, 2005/0048512, and 2004/0175756, all of which are incorporated herein by reference herein whole).

The synthesis of avimers involves phage display libraries derived from the human domain repertoire

A. Synthetic recombination is used to create a very diverse pool of monomers, as described in Silverman et al. (2005 Nature Biotechnology 23: 1556-1561). After the generation of a monomer reserve, the reserve is screened for a target protein (e.g., human serum albumin). The initial candidates are identified, and an additional monomer is added and the resulting dimer library is screened for the target protein to identify candidate target dimer dimers. Next, the method for obtaining a trimer with very high binding affinity for the target protein is iterated, and, optionally, can be further iterated to identify higher order candidate complexes. The candidate complexes that are identified for binding with high affinity and specificity to target proteins are called avimers (by "avidity multimer").

The monomeric domains of avimers can be chains of polypeptides of any size. For example, monomeric domains may have from about 25 to about 500, from about 30 to about 200, from about 30 to about 100, from about 90 to about 200, from about 30 to about 250, from about 30 to about 60, from about 9 to about 150, from about 100 to about 150, from about 25 to about 50 or from about 30 to about 150 amino acids. Similarly, a monomer domain of an avimer may comprise, e.g. eg, from about 30 to about 200 amino acids; from about 25 to about 180 amino acids; from about 40 to about 150 amino acids; from about 50 to about 130 amino acids; or from about 75 to about 125 amino acids. Monomeric domains and immunodomains can normally maintain a stable conformation in solution. Occasionally, the monomeric domains of avimers and immunodomins can be folded independently in a stable conformation. The stable conformation can be stabilized by metal ions. The stable conformation may optionally contain disulfide bonds (eg, at least one, two, or three or more disulfide bonds). Disulfide bonds can optionally be formed between two cysteine residues.

Publications describing monomeric domains and mosaic proteins and references cited therein include the following: Hegyi, H and Bork, P. 1997 J. Protein Chem., 16: 545-551; Baron et al. 1991 Trends Biochem. Sci. 16: 13-17; Ponting et al. 2000 Adv. Protein Chem. 54: 185-244; Doolittle 1995 Annu. Rev. Biochem 64: 287-314; Doolitte and Bork 1993 Scientific American 269: 50-6; and Bork 1991 FEBS Letters 286: 47-54. Monomeric domains used in avimers may also include the domains present in the Pfam database and the SMART database. See Schultz et al. 2000 Nucleid Acid Res. 28: 231-34.

Monomeric domains that are especially suitable for use in avimer compositions are (1) sandwich domains; (2) β domains in barrel; or (3) cysteine rich domains comprising disulfide bonds. The cysteine-rich domains used in avimers do not normally form a structure of α, β-blade or β-barrel helix. Normally, disulfide bonds promote domain folding in a three-dimensional structure. Normally, cysteine-rich domains have at least two disulfide bonds, more usually at least three disulfide bonds.

Monomeric domains of avimers can have any number of characteristics. For example, domains may have low or non-existent immunogenicity in an animal (eg, a human being). The domains may have a small size, for example, the domains may be small enough to penetrate the skin or other tissues. Domains may have a range of half-lives or in vivo stabilities.

Illustrative monomer domains suitable for use in avimer compositions include, e.g. eg, an EGF type domain, a Kringle domain, a fibronectin type I domain, a fibronectin type II domain, a fibronectin type III domain, a PAN domain, a Gla domain, an SRCR domain, a Kunitz / bovine pancreatic trypsin inhibitor domain, a Kazal-type serine protease inhibitor domain, a (type P) clover domain, a C-type domain of von Willebrand factor, an anaphylatoxin-like domain, a CUB domain, a type I repeat of thyroglobulin, an LDL receptor class domain, a Sushi domain, a Link domain, a type I thrombospondin domain, a thyromlobulin monomer domain, an immunoglobulin type domain, a lectin domain of type C, an MAM domain, a von Willebrand factor type domain, a somatomedin B domain, a central domain of four WAP type disulfides, a F5 / 8 type C domain, a hemopexin domain , a SH2 domain, an SH3 domain, a domain of type Laminin-type EFG, a C2 domain and other domains known to those skilled in the art, as well as derivatives and / or variants thereof. U.S. Patent Publication No. 20050221384 presents schematic diagrams of various example forms of monomeric domains present in the molecules in the LDL receptor family.

Suitable monomeric domains (e.g., domains with the ability to fold independently or with limited assistance) can be selected from families of protein domains that contain three-dimensional structures sandwich β or barrel β as defined by analysis tools. computational sequences such as Simple Modular Architecture Research Tool (SMART; see Shultz et al. 2000 Nucleic Acids Research 28: 231-234) or CATH (see Pearl et al. 2000 Nucleic Acids Research 28: 277-282). The monomeric domains of example avimers also include fibronectin type III domains, an anticalin domain and an Ig type domain of CTLA-4. Some aspects of these domains are described in WO01 / 64942 of Lipovsek et al., WO99 / 16873 of Beste et al. and WO-00/60070 of Desmet et al.

The monomeric domains of avimers are optionally rich in cysteine. Monomeric cysteine-rich domains include, for example, the class A LDL receptor domain ("domain A") or the type EGF domain. Monomeric domains can also have a group of negatively charged residues. Optionally, the monomeric domains contain a repeated sequence, such as YWTD as found in the β propellant domain. Another example monomer domain suitable for use in avimers is the C2 domain. Example sequences of domain A and domain C2 and consensus sequences useful in the production of avimers are presented, including the example selections of amino acid residues (eg, loop residues exposed to the surface) most desirable for addressing mutagenic, in US Pat. No. 2005/0221384.

Polynucleotides (also referred to as nucleic acids) encoding the monomeric domains are normally used to prepare monomeric domains by expression. Nucleic acids encoding monomer domains can be obtained from several different sources. Monomeric domain libraries can be prepared by expressing a plurality of different nucleic acids encoding naturally occurring monomeric domains, altered monomeric domains (i.e., monomeric domain variants), or a combination thereof.

Monomeric domains that bind to a selected or desired ligand (e.g., human serum albumin) or mixture of ligands are optionally identified as an initial stage in the production of avimers. In some embodiments, the monomeric and / or immunodomain domains are identified or selected for a desired property (e.g., binding affinity for human serum albumin) and then the monomeric and / or immunodomain domains are formed into multimers. For such embodiments any method that produces the selection of domains with a desired property (eg, human serum albumin binding) can be used. For example, the methods may comprise the delivery of a plurality of different nucleic acids, each nucleic acid encoding a monomeric domain; translating the plurality of different nucleic acids, thus providing a plurality of different monomeric domains; screening the plurality of different monomer domains for ligand binding or mixing of desired ligands; and identifying the members of the plurality of different monomeric domains that bind to the ligand or mixture of desired ligands.

Monomeric domains for avimer production can be naturally occurring or altered (unnatural variants). The term "natural occurrence" is used herein to indicate that an object can be found in nature. For example, natural monomer domains may include human monomer domains or optionally, domains derived from different species or sources, e.g. eg, mammals, primates, rodents, fish, birds, reptiles, plants, etc. Monomeric domains of natural occurrence can be obtained by various methods, e.g. eg, by PCR amplification of genomic DNA or cDNA. The term "natural", as used herein, is used in reference to a nucleic acid and / or polypeptide that has not been altered by mutagenesis or other techniques by performing any of the methods described below.

The monomeric domains of avimers may be domains of natural occurrence or variants of unnatural occurrence. The libraries of monomeric domains used in the synthesis of avimers may contain monomeric domain of natural occurrence, variants of monomeric domains of unnatural occurrence or a combination thereof.

A variety of vectors or information presentation systems can be used to express nucleic acids encoding monomeric and avimeric domains, and to test a desired activity (eg, human serum albumin binding). For example, a phage display system is a system in which monomeric domains are expressed as phage surface fusion proteins (Pharmacia, Milwaukee Wis). Phage display may involve the presentation of a polypeptide sequence encoding monomeric domains and / or immunodomains on the surface of filamentous bacteriophages, usually as fusion with a bacteriophage coating protein. Example methods of affinity enrichment and phage display are indicated, for example, in PCT Patent Publications No. 91/17271, 91/18980 and 91/19818 and 93/08278.

Examples of other presentation systems include ribosome presentations, a nucleotide-linked presentation (see, e.g., U.S. Patent Nos. 6,281,344; 6,194,550, 6,207,446, 6,214,553 and 6,258 .558), cell surface presentations and the like. The cell surface presentations include various cells,

p. eg, E. coli, yeasts and / or mammalian cells. When a cell is used as a presentation, nucleic acids, e.g. eg, obtained by PCR amplification followed by digestion, are introduced into the cell and translated. Optionally, polypeptides encoding monomeric or avimeric domains can be introduced, e.g. eg, by injection, in the cell.

As described below and in the art, avimers are multimeric compositions. In exemplary embodiments, the multimers comprise at least two monomeric domains and / or immunodomains. For example, the multimers may comprise from 2 to about 10 monomeric and / or immunodomain domains, from 2 to about 8 monomeric and / or immunodomain domains, from about 3 and about 10 monomeric and / or immunodomain domains, about 7 monomeric domains and / or or immunodomains, approximately 6 monomeric and / or immunodomain domains, approximately 5 monomeric and / or immunodomain domains, or approximately 4 monomeric and / or immunodomain domains. In some embodiments, the multimer comprises at least 3 monomeric domains and / or immunodomains. Normally, monomeric domains have been preselected for binding to the target molecule of interest (eg, human serum albumin).

Within an avimer, each monomer domain can specifically bind to a target molecule (eg, human serum albumin). Optionally, each monomer is attached in a different position (analogously to an epitope) in a target molecule. The binding of multiple monomeric and / or immunodomain domains to the same target molecule can produce an avidity effect that results in an avidity enhancement of the multimeric avimer to the target molecule compared to each individual monomer. Optionally, the multimer may have an avidity of at least about 1.5, 2, 3, 4, 5, 10, 20, 50 or 100 times the avidity of a single monomer domain for the target protein (e.g., human serum albumin).

The selected monomer domains can be linked by a linker group to form a multimer (avimer). For example, a linker group is placed between each discrete monomer domain separated in a multimer. Normally, the immunodomains are also linked to each other or to monomeric domains by means of a linker group moiety. The moieties of linker groups that can be easily used to bind immunodomain variants together are the same as those described for monomer domain multimer variants. Examples of linker groups suitable for the binding of immunodomain variants to other domains in multimers are described herein.

The binding of monomeric domains selected by a linker group to form an avimer can be accomplished using a variety of techniques known in the art. For example, the combinatorial set of polynucleotides encoding selected monomer domains can be achieved by DNA ligation, or optionally, by self-priming reactions based on PCR. The linker group may be attached to a monomer before identifying the monomer in terms of its ability to bind to a target multimer or after the monomer has been selected for its ability to bind to a target multimer.

As mentioned above, the polypeptide (s) comprising avimers can be altered. Descriptions of a variety of various generation procedures for the generation of modified or altered nucleic acid sequences encoding these polypeptides have been described above and are described below in the following publications and references cited therein: Soong, N et al., Molecular breeding of viruses, (2000) Nat Genet 25 (4): 436-439; Stemmer, et al., Molecular breeding of viruses for targeting and other clinical properties, (1999) Tumor Targeting 4: 1-4; Ness et al., DNA Shuffling of subgenomic sequences of subtilisin, (1999) Nature Biotechnology 17: 893-896; Chang et al., Evolution of a cytokine using DNA family shuffling, (1999) Nature Biotechnology 17: 793-797; Minshull and Stemmer, Protein evolution by molecular breeding, (1999) Current Opinion in Chemical Biology 3: 284-290; Christians et al., Directed evolution of thymidine kinase for AZT phosphorylation using DNA family shuffling, (1999) Nature Biotechnology 17: 259-264; Crameri et al., DNA shuffling of a family of genes from diverse species accelerates directed evolution, (1998) Nature 391: 288-291; Crameri et al., Molecular evolution of an arsenate detoxification pathway by DNA shuffling, (1997) Nature Biotechnology 15: 436-438; Zhang et al., Directed evolution of an effective fucosidase from a galactosidase by DNA shuffling and screening (1997) Proc. Natl Acad. Sci. USA 94: 4504-4509; Patten et al., Applications of DNA Shuffling to Pharmaceuticals and Vaccines, (1997) Current Opinion in Biotechnology 8: 724-733; Crameri et al., Construction and evolution of antibody-phage libraries by DNA shuffling, (1996) Nature Medicine 2: 100-103; Crameri et al., Improved green fluorescent protein by molecular evolution using DNA shuffling, (1996) Nature Biotechnology 14: 315319; Gates et al., Affinity selective isolation of ligands from peptide libraries through display on a lac repressor ’headpiece dimer’, (1996) Journal of Molecular Biology 255: 373-386; Stemmer, Sexual PCR and Assembly PCR, (1996) In: The Encyclopedia of Molecular Biology. VCH Publishers, New York. pp. 447-457; Crameri and Stemmer, Combinatorial multiple cassette mutagenesis creates all the permutations of mutant and wildtype cassettes, (1995) BioTechniques 18: 194-195; Stemmer et al., Single-step assembly of a gene and entire plasmid form large numbers of oligodeoxy-ribonucleotides, (1995) Gene, 164: 49-53; Stemmer, The Evolution of Molecular Computation, (1995) Science 270: 1510; Stemmer Searching Sequence Space, (1995) Bio / Technology 13: 549-553; Stemmer, Rapid evolution of a protein in vitro by DNA shuffling, (1994) Nature 370: 389-391; and Stemmer, DNA shuffling by random fragmentation and reassembly: In vitro recombination for molecular evolution, (1994) Proc. Natl Acad. Sci. USA 91: 10747-10751.

Methods of mutations for diversity generation include, for example, site-directed mutagenesis (Ling et al., Approaches to DNA mutagenesis: an overview, (1997) Anal Biochem. 254 (2): 157-178; Dale et al ., Oligonucleotide-directed random mutagenesis using the phosphorothioate method, (1996) Methods Mol. Biol. 57: 369-374; Smith, In vitro mutagenesis, (1985) Ann. Rev. Genet. 19: 423-462; Botstein & Shortle , Strategies and applications of in vitro mutagenesis, (1985) Science 229: 1193-1201; Carter, Site-directed mutagenesis, (1986) Biochem. J. 237: 1-7; and Kunkel, The efficiency of oligonucleotide directed mutagenesis, ( 1987) in Nucleic Acids & Molecular Biology (Eckstein, F. and Lilley, DMJ eds., Springer Verlag, Berlin)); mutagenesis using uracil containing templates (Kunkel, Rapid and efficient site-specific mutagenesis without phenotypic selection, (1985) Proc. Natl. Acad. Sci. USA 82: 488-492; Kunkel et al., Rapid and efficient site-specific mutagenesis without phenotypic selection, (1987) Methods in Enzymol. 154, 367-382; and Bass et al., Mutant Trp repressors with new DNA-binding specificities, (1988) Science 242: 240-245); oligonucleotide-directed mutagenesis ((1983) Methods in Enzymol. 100: 468-500; (1987) Methods in Enzymol. 154: 329-350; Zoller & Smith, Oligonucleotide-directed mutagenesis using M13-derived vectors: an efficient and general procedure for the production of point mutations in any DNA fragment, (1982) Nucleic Acids Res. 10: 6487-6500; Zoller & Smith, Oligonucleotide-directed mutagenesis of DNA fragments cloned into M13 vectors, (1983) Methods in Enzymol. 100: 468 -500; and Zoller & Smith, Oligonucleotide-directed mutagenesis: a simple method using two oligonucleotide primers and a single-stranded DNA template, (1987) Methods in Enzymol. 154: 329-350); phosphorothioate-modified DNA mutagenesis (Taylor et al., The use of phosphorothioate-modified DNA in restriction enzyme reactions to prepare nicked DNA, (1985) Nucl. Acids Res. 13: 8749-8764; Taylor et al., The rapid generation of oligonucleotide-directed mutations at high frequency using phosphorothioate-modified DNA, (1985) Nucl. Acids Res. 13: 8765-8787; Nakamaye & Eckstein, Inhibition of restriction endonuclease Nci I cleavage by phosphorothioate groups and its application to oligonucleotide-directed mutagenesis, (1986) Nucl. Acids Res. 14: 9679-9698; Sayers et al., YT Exonucleases in phosphorothioate-based oligonucleotide-directed mutagenesis, (1988) Nucl. Acids Res. 16: 791-802; and Sayers et al., Strand specific cleavage of phosphorothioate-containing DNA by reaction with restriction endonucleases in the presence of ethidium bromide, (1988) Nucl. Acids Res. 16: 803-814); mutagenesis using gapped duplex DNA (Kramer et al., The gapped duplex DNA approach to oligonucleotide-directed mutation construction, (1984) Nucl. Acids Res. 12: 9441-9456; Kramer & Fritz Oligonucleotide-directed construction of mutations via gapped duplex DNA , (1987) Methods in Enzymol. 154: 350-367; Kramer et al., Improved enzymatic in vitro reactions in the gapped duplex DNA approach to oligonucleotide-directed construction of mutations, (1988) Nucl. Acids Res. 16: 7207; and Fritz et al., Oligonucleotidedirected construction of mutations: a gapped duplex DNA procedure without enzymatic reactions in vitro, (1988) Nucl. Acids Res. 16: 6987-6999).

Additional suitable methods include repair of absence of point pairing (Kramer et al., Point Mismatch Repair, (1984) Cell 38: 879-887), mutagenesis using repair-deficient host strains (Carter et al., Improved oligonucleotide site-directed mutagenesis using M13 vectors, (1985) Nucl. Acids Res. 13: 4431-4443; and Carter, Improved oligonucleotide-directed mutagenesis using M13 vectors, (1987) Methods in Enzymol. 154: 382403), deletion mutagenesis (Eghtedarzadeh & Henikoff, Use of oligonucleotides to generate large deletions, (1986) Nucl. Acids Res. 14: 5115), restriction-selection and restriction-purification (Wells et al., Importance of hydrogen-bond formation in stabilizing the transition state of subtilisin, (1986 ) Phil. Trans. R. Soc. Lond. A 317: 415-423), mutagenesis by total gene synthesis (Nambiar et al., Total synthesis and cloning of a gene coding for the ribonuclease S protein, (1984) Science 223: 1299-1301; Sakamar and Khorana, Total synthesis and express ion of a gene for the a-subunit of bovine rod outer segment guanine nucleotide-binding protein (transduction), (1988) Nucl. Acids Res. 14: 6361-6372; Wells et al., Cassette mutagenesis: an efficient method for generation of multiple mutations at defined sites, (1985) Gene 34: 315-323; and Grundstrom et al., Oligonucleotide-directed mutagenesis by microscale ’shot-gun’ gene synthesis, (1985) Nucl. Acids Res. 13: 3305-3316), double-strand break repair (Mandecki, Oligonucleotide-directed double-strand break repair in plasmids of Escherichia coli: a method for sitespecific mutagenesis, (1986) Proc. Natl. Acad. Sci. USA , 83: 7177-7181; and Arnold, Protein engineering for unusual environments, (1993) Current Opinion in Biotechnology 4: 450-455). Additional details of many of the above methods can be found in Methods in Enzymology Volume 154, which also describes useful controls for repair problems with various mutagenesis methods.

Additional details regarding various methods of generating diversity can be found in the following US patents, PCT publications and applications and EPO publications: US patent. nº

5,605,793 to Stemmer (Feb. 25, 1997), "Methods for In Vitro Recombination"; US Patent No. 5,811,238 to Stemmer et al. (Sep. 22, 1998) "Methods for Generating Polynucleotides having Desired Characteristics by Iterative Selection and Recombination ”; U.S. Patent No. 5,830,721 for Stemmer et al. (Nov. 3, 1998), "DNA Mutagenesis by Random Fragmentation and Reassembly"; US Patent No. 5,834,252 to Stemmer, et al. (Nov. 10, 1998) "End-Complementary Polymerase Reaction"; U.S. Patent No. 5,837,458 to Minshull, et al. (Nov. 17, 1998), "Methods and Compositions for Cellular and Metabolic Engineering"; WO 95/22625, Stemmer and Crameri, "Mutagenesis by Random Fragmentation and Reassembly"; WO 96/33207 for Stemmer and Lipschutz "End Complementary Polymerase Chain Reaction"; WO 97/20078 for Stemmer and Crameri "Methods for Generating Polynucleotides having Desired Characteristics by Iterative Selection and Recombination"; WO 97/35966 for Minshull and Stemmer, "Methods and Compositions for Cellular and Metabolic Engineering"; WO 99/41402 for Punnonen et al. "Targeting of Genetic Vaccine Vectors"; WO 99/41383 for Punnonen et al. "Antigen Library Immunization"; WO 99/41369 for Punnonen et al. "Genetic Vaccine Vector Engineering"; WO 99/41368 for Punnonen et al. "Optimization of Immunomodulatory Properties of Genetic Vaccines”; EP 752008 for Stemmer and Crameri, "DNA Mutagenesis by Random Fragmentation and Reassembly”; EP 0932670 for Stemmer "Evolving Cellular DNA Uptake by Recursive Sequence Recombination"; WO 99/23107 for Stemmer et al., "Modification of Virus Tropism and Host Range by Viral Genome Shuffling"; WO 99/21979 for Apt et al., "Human Papillomavirus Vectors"; WO 98/31837 by del Cardayre et al. "Evolution of Whole Cells and Organisms by Recursive Sequence Recombination"; WO 98/27230 for Patten and Stemmer, "Methods and Compositions for Polypeptide Engineering"; WO 98/27230 for Stemmer et al., "Methods for Optimization of Gene Therapy by Recursive Sequence Shuffling and Selection," WO 00/00632, "Methods for Generating Highly Diverse Libraries," WO 00/09679, "Methods for Obtaining in Vitro Recombined Polynucleotide Sequence Banks and Resulting Sequences," WO 98/42832 for Arnold et al., "Recombination of Polynucleotide Sequences Using Random or Defined Primers, "WO 99/29902 for Arnold et al.," Method for Creating Polynucleotide and Polypeptide Sequences, "WO 98/41653 for Vind," An in Vitro Method for Construction of a DNA Library, "WO 98/41622 for Borchert et al. , "Method for Constructing a Library Using DNA Shuffling," and WO 98/42727 for Pati and Zarling, "Sequence Alterations using Homologous Recombination”; WO 00/18906 for Patten et al., "Shuffling of Codon-Altered Genes"; WO 00/04190 for Cardayre et al. "Evolution of Whole Cells and Organisms by Recursive Recombination"; WO 00/42561 for Crameri et al., "Oligonucleotide Mediated Nucleic Acid Recombination"; WO 00/42559 for Selifonov and Stemmer "Methods of Populating Data Structures for Use in Evolutionary Simulations"; WO 00/42560 for Selifonov et al., "Methods for Making Character Strings, Polynucleotides & Polypeptides Having Desired Characteristics"; WO 01/23401 for Welch et al., "Use of Codon-Varied Oligonucleotide Synthesis for Synthetic Shuffling"; and PCT / US01 / 06775 "Single-Stranded Nucleic Acid Template-Mediated Recombination and Nucleic Acid Fragment Isolation" for Affholter.

Polypeptides (eg, avimers) can optionally be expressed in cells. Multimer domains can be synthesized as a single protein using expression systems well known in the art. In addition to the numerous texts indicated above, the general texts describing molecular biology techniques useful herein, including the use of vectors, promoters and many other topics relevant to the expression of nucleic acids such as monomeric domains, selected monomeric domains, selected multimers and / or multimers, include Berger and Kimmel, Guide to Molecular Cloning Techniques, Methods in Enzymology volume 152 Academic Press, Inc., San Diego, Calif. (Berger); Sambrook et al., Molecular Cloning - A Laboratory Manual (2nd Ed.), Vol. 1-3, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989 ("Sambrook") and Current Protocols in Molecular Biology, FM Ausubel et al., eds., Current Protocols, a joint venture between Greene Publishing Associates, Inc. and John Wiley & Sons, Inc., (supplemented through 1999) ("Ausubel")). Examples of sufficient techniques are found to direct those skilled in the art through in vitro amplification methods, useful for identifying the isolation and cloning of domains of monomers and multimers encoding nucleic acids, which include the chain reaction of the polymerase (PCR), ligase chain reaction (CSF), amplification of replicase Q and other RNA-polymerase-mediated techniques (e.g., NASBA), in Berger, Sambrook, and Ausubel, as well as in Mullis et al., (1987) US Pat. No. 4,683,202; PCR Protocols A Guide to Methods and Applications (Innis et al. Eds) Academic Press Inc. San Diego, Calif. (1990) (Innis); Arnheim & Levinson (Oct. 1, 1990) C&EN 36-47; The Journal Of NIH Research (1991) 3, 81-94; (Kwoh et al. (1989) Proc. Natl. Acad. Sci. USA 86, 1173; Guatelli et al. (1990) Proc. Natl. Acad. Sci. USA 87, 1874; Lomell et al. (1989) J. Clin. Chem 35, 1826; Landegren et al., (1988) Science 241, 1077-1080; Van Brunt (1990) Biotechnology 8, 291-294; Wu and Wallace, (1989) Gene 4, 560; Barringer et al. (1990) Gene 89, 117, and Sooknanan and Malek (1995) Biotechnology 13: 563-564. Enhanced methods of cloning of in vitro amplified nucleic acids are described in Wallace et al., U.S. Patent No. 5,426. 039. The improved methods of amplification of large nucleic acids by PCR are summarized in Cheng et al. (1994) Nature 369: 684-685 and references thereof, in which PCR amplicons of up to 40 kb are generated. you will see that essentially any RNA can be converted into a double stranded DNA suitable for restriction digestion, PCR expansion and sequencing using reverse transcriptase and a polymerase.

Encoding vectors can be introduced, e.g. eg, monomeric and / or avimeric domains in host cells, produced and / or selected by recombinant techniques. Host cells are designed (ie, transduced, transformed or transfected) with said vectors, which can be, for example, a cloning vector or an expression vector. The vector may be, for example, in the form of a plasmid, a viral particle, a phage, etc. Designed host cells can be cultured in conventional nutrient media modified as appropriate to activate promoters, select transformants or amplify the gene or genes of the monomer domain, the selected monomer domain, the multimer and / or the selected multimer of interest. Culture conditions, such as temperature, pH and the like, are those previously used with the host cell selected for expression, and will be apparent to those skilled in the art and in the references cited herein, including, e.g. eg Freshney (1994) Culture of Animal Cells, a Manual of Basic Technique, 3rd edition, Wiley-Liss, New York and the references cited therein.

Polypeptides can also be produced in non-animal cells such as plants, yeasts, fungi, bacteria and the like. In fact, as noted above, phage display is an especially relevant technique for producing such polypeptides. In addition to Sambrook, Berger and Ausubel, details regarding cell culture can be found in Payne et al. (1992) Plant Cell and Tissue Culture in Liquid Systems John Wiley & Sons, Inc. New York, N.Y .; Gamborg and Phillips (eds) (1995) Plant Cell, Tissue and Organ Culture; Fundamental Methods Springer Lab Manual, Springer-Verlag (Berlin Heidelberg New York) and Atlas and Parks (eds) The Handbook of Microbiological Media (1993) CRC Press, Boca Raton, Fla.

Avimers may possess alterations of monomeric, immunodomain and / or multimeric domains that improve pharmacological properties, reduce immunogenicity or facilitate transport of the multimer and / or the monomeric domain in a cell or tissue (e.g., through the blood brain barrier, or through the skin). These types of alterations include various modifications (e.g., addition of sugar groups or glycosylation), the addition of PEG, the addition of protein domains that bind to a certain protein (e.g., HSA or other protein serum), the addition of protein fragments or sequences that signal movement or transport to, from and through a cell. Additional components may also be added to a multimer and / or monomer domain to manipulate the properties of the multimer and / or monomer domain. Various components may also be added including, e.g. eg, a domain that binds to a known receptor (e.g., an Fc region protein domain that binds to an Fc receptor), one or more toxins or part of a toxin, a dominance that can optionally be cleaved to activate the multimer or monomer domain, a reporter molecule (e.g., green fluorescent protein), a component that binds to the reporter molecule (such as a radionuclide for radiotherapy, biotin or avidin) or a combination of modifications.

As used herein, "directed evolution" refers to a process by which polynucleotide variants are generated, expressed or screened for an activity (eg, a polypeptide with binding activity for a target protein of target human serum albumin) in a recursive process. One or more candidates are selected in the screening and then the process is repeated using polynucleotides encoding the selected candidates to generate new variants. Directed evolution involves at least two batches of variation generation and may include 3, 4, 5, 10, 20 or more batches of variation and selection generation. The variation can be generated by any method known to those skilled in the art, which includes, e.g. eg, error-prone PCR, gene shuffling, chemical mutagenesis and the like.

The term "shuffled" is used herein to indicate a recombination between non-identical sequences. In some embodiments, shuffling may include crossover by means of homologous or non-homologous recombination, for example, by means of cre / lox and / or flp / frt systems. Shuffling can be done using a variety of different formats, including, for example, in vitro and in vivo shuffling formats, in silico shuffling formats, shuffling formats using double-stranded or single-stranded molds, primer-based shuffle formats, shuffled based on nucleic acid fragmentation and oligonucleotide-mediated shuffle formats, all of which are based on recombination episodes between non-identical sequences and are described in more detail or referred to herein herein, as well as others similar formats based on recombination.

The term "random" as used herein refers to a polynucleotide sequence or an amino acid sequence composed of two or more amino acids and constructed by a stochastic or random process. The polynucleotide sequence or random amino acid sequence may include frame or framework motifs, which may comprise invariant sequences.

GroEL and GroES

Example 24: Generation of a double specific ligand comprising a cpn10 non-immunoglobulin (GroES) framework of CDR graft medium serum albumin binding

The CDR3 domain of dAb7h14 is used to construct a cpn10 non-immunoglobulin scaffold polypeptide (GroES) that binds to human serum albumin in the following manner. The CDR3 sequence (AQGAALPRT; SEQ ID NO.:__) of dAb7h14 is grafted onto the cpn10 polypeptide replacing natural cpn10 amino acid residues at positions 19-27 (movable loop residues). Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to immobilized cpn10 derived polypeptide comprising the human serum anti-albumin CDR domain of dAb7h14. (One skilled in the art will recognize that binding affinity can be assessed using any appropriate method, including, e.g., precipitation of labeled human serum albumin, competitive Biacore assay, etc.). If a low or no human serum albumin affinity is detected (e.g., Kd values in the range of M or higher), at least one of a set of strategies is used to improve serum albumin binding properties human CDR3 grafted cpn10 polypeptide, which includes any of the following methods that contribute to binding affinity.

The length of the CD3 grafted region of dAb7h14 of the cpn10 polypeptide corresponding to the mobile loop region in natural cpn10 polypeptides is adjusted through the deletion of amino acid residues and / or the use of linker group polypeptides. For example, the CDR2 peptide sequence of nine amino acid residues of dAb7h14 extends to 16 amino acid residues in length using amino acid linker groups of, e.g. eg, zero to seven length residues located on the flanks of the N, C or both ends of the CDR dAb7h14 polypeptide sequence, thus achieving a total length of the 16 amino acid grafted peptide sequence in the CDR3 grafted domain corresponding to the mobile loop in the native cpn10 sequence. Such use of polypeptide or linker group polypeptides is optionally combined with mutagenesis of linker group sequences, sequence or sequences of CDR3 and / or cpn10 sequences of non-CDR (eg, using mutagenic optimization procedures as described further) below), in order to improve the ability of human serum albumin binding of cpn10 polypeptides grafted with CDR3 (e.g., by optimization of the CDR and fibronectin sequences in cpn10 polypeptides grafted with CDR3). The polypeptide linker groups used for this purpose have a predetermined sequence, or, optionally, are selected from a population of polypeptide sequences of randomized linker groups by evaluation of the binding serum human albumin binding capabilities containing cpn10 polypeptides grafted with CDR3. Optimization methods are performed in parallel and / or iteratively. Both parallel and iterative optimization processes (e.g., affinity maturation) employ selection methods as described below and / or as are known in the art useful for optimizing polypeptide binding properties. .

Human serum albumin binding of polypeptides or CDR-grafted polypeptides presenting CDR3 dAb7h14 is optimized by mutagenesis, optionally in combination with parallel and / or iterative selection methods as described below and / or known by other means in the technique. The cpn10 framework polypeptide domains surrounding the grafted CDR3 dAb7h14 polypeptide sequence are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed in the cpn10 polypeptide sequence on non-grafted amino acid residues, and is optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. Optionally, the domain of CDR3 dAb7h14 polypeptides presented in the CDR3 grafted cpn10 polypeptide is subjected to mutagenesis by, e.g. eg, random mutagenesis, NNK mutagenesis, review mutagenesis and / or other method recognized by the art. PCR is optionally used to perform such mutagenesis methods, which results in the generation of a variety of sequences through targeted sequences in cpn10 polypeptides grafted with CDR3. These strategies are similar to those described below for generating dAb libraries. In addition to methods of random and / or review mutagenesis, directed mutagenesis of targeted amino acid residues is used where the structural information establishes specific amino acid residues that are essential for human serum albumin binding.

The cpn10 polypeptides comprising grafted CDR3 dAb7h14 CDR3 sequence designed as described above are subjected to parallel and / or iterative selection methods to identify those cpn10 polypeptides that are optimized for human serum albumin binding. For example, after the production of a library of cpn10 polypeptide sequences grafted with CDR3 dAb7h14, this library of such polypeptides is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described later for selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized monomeric and / or oligomeric cpn10 derived polypeptides that ideally they achieve Kd values of serum albumin binding affinity in the range of nM or better.

Upon identification of monomeric cpn10 derived polypeptides that bind human serum albumin, the human serum binding properties of said initial monomers can be further enhanced by combining said monomers with other monomers, followed by mutagenesis and / or additional selection , thereby forming an oligomeric composition cpn10 / GroES having specific affinity for human serum albumin. After the identification of a cpn10 / GroES oligomeric composition having human serum albumin affinity, said polypeptides are then used to generate double specific ligand compositions by any of the methods described below.

Example 25: Generation of a double specific ligand comprising a cpn10 non-immunoglobulin framework for serum albumin binding through the selection of serum albumin binding moieties

The natural cpn10 polypeptide is subjected to library selection techniques and, optionally, affinity maturation in order to produce cpn10 non-immunoglobulin sheath molecules bound to serum albumin for use in specific double ligands.

The ability of a natural cpn10 polypeptide to bind to human serum albumin is initially determined by Biacore assay as described above. After the detection of null or low binding affinity (e.g., Kd values in the range of M or higher) of a cpn10 polypeptide for human serum albumin, at least one of a set of strategies is employed for imparting the human serum albumin binding properties to the cpn10 polypeptide, which includes one or more of the following methods that contribute to binding affinity.

Human serum albumin binding of cpn10 framework polypeptide or polypeptides is achieved and optimized by mutagenic methods, optionally in combination with parallel and / or iterative selection methods as described below and / or known by other means in the technique. The cpn10 framework polypeptide domains are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is carried out on the entire cpn10 polypeptide or on specific sequences in the cpn10 polypeptide, which includes residues of mobile amino acid bonds at positions 19-27, and is optionally randomized in order to obtain new polypeptides by evolution. Enhanced human serum albumin binding. PCR is optionally used to perform said mutagenesis methods, which results in the generation of a variety of sequences through sequences directed at cpn10 polypeptides. (These strategies are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is used where the structural information establishes specific amino acid residues of cpn10 polypeptides that are critical for the binding of human serum albumin.

The cpn10 polypeptides designed as described above are subjected to parallel and / or iterative selection methods to identify those cpn10 polypeptides that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized cpn10 polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for selection. of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized monomeric and / or oligomeric cpn10 derived polypeptides that ideally they reach Kd values of serum albumin binding affinity in the range of nM or better.

Upon identification of monomeric cpn10 derived polypeptides that bind human serum albumin, the human serum binding properties of said initial monomers can be further enhanced by combining said monomers with other monomers, followed by mutagenesis and / or additional selection , thus forming an oligomeric cpn10 / GroES composition that has specific affinity for human serum albumin. After identification of an oligomeric cpn10 / GroES composition that has an affinity for human serum albumin, said polypeptides are used to generate double specific ligand compositions by any of the methods described below.

GroEL polypeptides

GroEL is a key molecular chaperone in E. coli consisting of 14 subunits each of a molecular mass of 57.5 Kd arranged in two seven-link rings (Braig et al. 1994 Proc. Natl. Acad. Sci. 90: 39783982 ). There is a large cavity in the GroEL ring system, and as widely believed the cavity is necessary for a successful protein folding activity. For optimal activity, a cochaperone, GroES, is required, consisting of a seven-link ring of 10 Kd subunits (Hunt et al. 1996 Nature

379: 37-45). Each GroES subunit uses a mobile loop with a conserved hydrophobic tripeptide for interaction with GroEL (Landry et al. 1993 Nature 364: 255-258). Mobile loops are generally less than 16 amino acids in length and undergo a transition from disordered loops to concomitant β hairpins with the union of GroEL's apical domains (Shewmaker et al. 2001 J. Biol. Chem. 276: 31257-31264) . The activity of the GroEL / GroES complex requires ATP. GroEL and GroES are widespread in all organisms, and are often referred to as chaperonin (cpn), cpn60 and cpn10 molecules, respectively.

GroEL is an allosteric protein. Allosteric proteins form a special class of oligomeric proteins, which alternate between two or more different three-dimensional structures during the binding of ligands and substrates. Allosteric proteins are often involved in control processes in biology or in which mechanical and physicochemical forms of energy are interconverted. The function of ATP is to trigger this allosteric change, causing the conversion of GroEL from a state that binds strongly denatured proteins to another that binds weakly denatured proteins. Co-chaperone, GroES, helps in this process by favoring the weak state of union. It can also act as a lid, closing the GroEL cavity. In addition, their binding to GroEL is likely directly to compete with the binding of denatured substrates. The net result is that the union of GroES and ATP to GroEL that has a substrate bound in its denatured form releases the denatured substrate in the cavity or in solution in which it can be folded back.

GroEL and GroES are polypeptide frameworks that can be used to multimerize polypeptides or domains of monomeric proteins, to produce multimeric proteins that have any desired characteristics. As also described below, p. For example, for avimer compositions, it is often desirable to multimerize polypeptide monomers.

Many proteins need the assistance of molecular chaperones to fold in vivo or fold back in vitro with high yields. Molecular chaperones are proteins, which are often large and need an energy source such as ATP to function. A key molecular chaperone in E. coli is GroEL, which consists of 14 subunits each of a 57.5 kDa molecular mass arranged in two seven-link rings. In the GroEL ring system there is a large cavity, and as it is widely believed the cavity is necessary for a successful folding activity. For optimal activity, a co-chaperone, GroES, is needed, consisting of a seven-link ring of 10 kDa subunits. The activity of the GroEL / GroES complex requires ATP as a source of energy.

The mini-chaperones have been described in detail elsewhere (see international patent application WO99 / 05163). Chaperone minipolypeptides have chaperone activity when they are in monomeric form and do not need energy in the form of ATP. Defined fragments of the apical domain of GroEL of approximately 143-186 amino acid residues in length have molecular chaperone activity towards proteins in solution in monomeric conditions or when they are monodispersed and bound to a support.

GroEL and / or GroES frameworks allow oligomerization of polypeptides to form functional protein oligomers that have activities that exceed those of recombinant monomeric polypeptides. Cpn10 is an extended component of the cpn60 / cpn10 chaperonin system. Examples of cpn10 include bacterial GroES and bacteriophage Gp31 T4, and are also listed below. Those skilled in the art will meet more members of the cpn10 family.

Protein shell subunits gather to form a protein shell. Said framework can have any shape and can comprise any number of subunits. For some embodiments of GroEL and GroES, the framework comprises between 2 and 20 subunits, between 5 and 15 subunits, or approximately 10 subunits. The natural occurrence framework structure of the members of the cpn10 family comprises seven subunits, in the form of a seven-link ring. In some embodiments, therefore, the frame is a seven-link ring.

A heterologous amino acid sequence can be inserted, which can be, e.g. eg, a CDR3 domain from an antibody or antigen-binding fragment thereof that has an affinity for a target protein (e.g., human serum albumin) or, optionally, which can be an individual residue such as cysteine that allows the binding of additional groups or molecules to the framework, in the sequence of the oligomerizable protein shell subunit so that the N and C ends of the polypeptide monomer are formed by the sequence of the oligomerizable protein shell subunit. Thus, the heterologous polypeptide is included with the subunit sequence of the framework, for example substituting one or more amino acids thereof.

It is known that cpn10 subunits have a "moving loop" in their structure. The mobile loop is located between amino acids 15 and 34, preferably between amino acids 16 to 33, of the GroES sequence of E. coli, and in equivalent positions in other members of the cpn10 family. The mobile loop of Gp31 T4 is located between residues 22 to 45, preferably 23 to 44. Optionally, the heterologous polypeptide can be inserted by replacing part or all of the moving loop of a polypeptide of the cpn10 family. When the subunit of the protein framework is a cpn10 family polypeptide, the heterologous sequence can be incorporated on the other hand at the N or C end thereof, or at positions that are equivalent to the roof fork b of the peptides of the cpn10 family. This position is located between positions 54 and 67, preferably 55 to 66, and preferably 59 and 61 of bacteriophage Gp31 T4, or between positions 43 to 63, preferably 44 to 62, advantageously 50 to 53 of E. coli GroES .

Optionally, a polypeptide can be inserted at the N or C end of a framework subunit in association with circular permutation of the subunit itself. Circular permutation is described in Graf and Schachman, PNAS (USA). 1996, 93: 11591. Essentially, the polypeptide is subjected to circular permutation by fusion of the existing N and C ends, and cleavage of the polypeptide chain elsewhere to create new N and C ends. In a preferred embodiment of the Invention, the heterologous polypeptide can be included at the N and / or C end formed after circular permutation. The site of formation of the new ends may be selected according to the desired characteristics, and may include the movable loop and / or the roof fork β.

Advantageously, heterologous sequences, which may be the same or different, can be inserted in more than one of the positions and / or in positions other than the positions previously identified within the subunit of the protein framework. Thus, each subunit may comprise two or more heterologous polypeptides, which are presented in the framework when it is assembled. Heterologous polypeptides may occur in a free or restricted subunit of the framework, depending on the degree of freedom provided by the insertion site in the framework sequence. For example, if the length of the sequences flanking the movable or β-hairpin loops in the framework is modified, the degree of restriction of any heterologous polypeptide inserted therein will be modulated.

The GroEL and / or GroES compositions may also comprise a polypeptide oligomer comprising two or more monomers. The oligomer may be configured as a heterooligomer, comprising two or more different amino acid sequences inserted into the framework, or as a homooligomer, in which the sequences inserted into the framework are the same.

The monomers that constitute the oligomer can be covalently crosslinked with each other. Crosslinking can be performed by recombinant approaches, so that the monomers are expressed ab initio as an oligomer; alternatively, crosslinking can be performed on Cys residues in the frame. For example, singular Cys residues inserted between positions 50 and 53 of the GroES framework, or equivalent positions in other members of the cpn10 family, can be used to cross-link the framework subunits.

The nature of the heterologous polypeptide inserted into the framework subunit can be selected as desired. In some embodiments, shell proteins that synthesize antibodies or fragments thereof such as scFv, natural or camelid VH domains and VR CDR fragments are synthesized.

In an exemplary embodiment, an oligomerization polypeptide monomer can be prepared as described above and / or as indicated in WO-00/69907, incorporated herein by reference in its entirety. The method of said preparation may comprise the insertion of a nucleic acid sequence encoding a heterologous polypeptide into a nucleic acid sequence encoding a subunit of an oligomerizable protein framework, which incorporates the resulting nucleic acid into an expression vector, and which expresses the nucleic acid to produce the polypeptide monomers. Optionally, a polypeptide oligomer can then be produced by means of a process comprising allowing the polypeptide monomers produced as before to be associated with an oligomer. In some embodiments, the monomers are crosslinked to form the oligomer.

In some embodiments, a framework polypeptide is based on members of the cpn10 / Hsp10 family, such as GroES or an analog thereof. A very preferred analog is the T4 Gp31 polypeptide. GroES analogs, which include Gp31, have a mobile loop (Hunt, JF, et al., (1997) Cell 90, 361-371; Landry, SJ, et al., (1996) Proc. Natl. Acad. Sci USA 93, 11622-11627) which can be inserted into, or replaced, in order to fuse the heterologous polypeptide into the framework.

The homologs of cpn10 are widespread in animals, plants and bacteria. For example, a GenBank search indicates that cpn10 homologs are known in the following species: Actinobacillus actinomycetemcomitans; Actinobacillus pleuropneumoniae; Aeromonas salmonicida; Agrobacterium tumefaciens; Allochromatium vinoum; Amoeba proteus symbiotic bacterium; Here / ex aeolicus; Arabidopsis thaliana; Bacillus sp; Bacillus stearothermophilus; Bacillus subtilis; Bartonella henselae; Bordetella pertussis; Borrelia burgdorferi; Brucella abortus; Buchnera aphidicola; Burkholderia cepacia; Burkholderia vietnamiensis; Campylobacterjejuni; Caulobacter crescentus; Chlamydia muridarum; Chlamydia trachomatis; Chlamydophila pneumoniae; Clostridium acetobutylicum; Clostridium perfringens; Clostridium thermocellum; T-Cowdria ruminantium coliphage; Cyanelle Cyanophora paradoxa; Ehrlichia canis; Ehrlichia chaffeensis; Ehrlichia equi; Ehrlichia phagocytophila; Ehrlichia risticii; Ehrlichia sennetsu; Ehrlichia sp HGE agent; Enterobacter aerogenes; Enterobacter agglomerans; Enterobacter amnigenus; Enterobacter asburiae; Enterobacter gergoviae; Enterobacter intermedius; Erwinia aphidicola; Erwinia carotovora; Erwinia herbicola; Escherichia coli; Francisella tularensis; Glycine max; Haemophilus ducreyi; Haemophilus influenzae Rd; Helicobacter pylori; Obtuse holospora; Homo sapiens; Klebsiella ornithinolytica; Klebsiella oxytoca; Klebsiella planticola; Klebsiella pneumoniae; Lactobacillus helvetictis; Lactobacillus Zeae; Lactococcus lactis; Lawsonia intracellularis; Leptospira interrogans; Methylovorus sp strain SS; Mycobacterium avium; Mycobacterium avium subsp avium; Mycobacterium avium subsp paratuberculosis; Mycobacterium leprae; Mycobacterium tuberculosis; Mycoplasma genitalium; Mycoplasma pneumoniae; Myzus persicae primary endosymbiont; Neisseria gonorrhoeae; Oscillatoria sp NKBG, -Pantoea ananas; Pasteurella multocida; Porphyromonas gingivalis; Pseudomonas aeruginosa; Pseudomonas aeruginosa; Pseudomonas putida; Rattus norvegicus; Rattus norvegicus; Rhizobium leguminosarum; Rhodobacter capsulatus; Rhodobacter sphaeroides; Rhodothermus marinus; Rickettsia prowazekii; Rickettsia rickettsii; Saccharomyces cerevisiae; Serratia ficaria; Serratia marcescens; Serratia rubidaea; Sinorhizobium meliloti; Sitophilus oryzae main endosymbiont; Stenotrophomonas maltophilia; Streptococcus pneumoniae; Streptomyces albus; Streptomyces coelicolor; Streptomyces coelicolor; Streptomyces lividans; Synechococcus sp; Synechococcus vulcanus; Synechocystis sp; Thermoanaerobacter brockii; Thermotoga maritima; Thermus aquaticus; Treponema pallidum; Wolbachia sp; Zymomonas mobilis.

An advantage of the subunits of the cpn10 family is that they have a mobile loop, responsible for the folding activity of the natural chaperonin protein, which can be eliminated without affecting the frame. The cpn10 with a suppressed mobile loop has no biological activity, which makes it an advantageously inert frame, thus minimizing any possible harmful effects.

The insertion of an appropriate biologically active polypeptide can confer a biological activity (eg, binding to human serum albumin) in the new polypeptide thus generated. In fact, the biological activity of the inserted polypeptide can be improved by incorporating the biologically active polypeptide into the framework, especially, e.g. eg, when affinity and mutagenesis based screening methods are used as described herein, they are used to optimize the binding to target protein of a polypeptide presented in the framework.

Alternative sites for peptide insertion are possible. An advantageous option is in the position equivalent to the roof fork β in GroES. This involves the replacement of Glu-in Gp31 with the desired peptide. The amino acid sequence is Pro (59) -Glu (60) -Gly (61). This sequence conveniently converts into a SmaI site at the DNA level (CCC: GGG) encoding Pro-Gly, leaving a blunt end insertion site for peptide insertion as a DNA fragment. Similarly, an insertion can be made between positions 50 and 53 of the GroES sequence, and in equivalent positions in other members of the cpn10 family. Alternatively, reverse PCR can be used to present the peptide on the opposite side of the framework.

Members of the cpn60 / Hsp60 family of chaperonin molecules can also be used as frameworks. For example, the GroEL tetradechameric bacterial chaperonin can be used. In some embodiments, heterologous polypeptides would be inserted between positions 191 and 376, in particular between positions 197 and 333 (represented by unique designed SacII and Cla I sites) to keep intact the hinge region between the equatorial and apical domains in order to impart mobility to the inserted polypeptide. The choice of the framework may depend on the intended application of the oligomer (or specific double ligand comprising and / or derived from said oligomer): for example, if the oligomer is intended for vaccination purposes, the use of an immunogenic framework, such as The derivative of Mycobacterium tuberculosis is highly advantageous and confers an adjuvant effect.

Mutants of cpn60 molecules can also be used. For example, the GroEL single ring mutant (GroELSRI) contains four point mutations that perform the main bond between the two GroEL rings (R452E, E461A, S463A and V464A) and is functionally inactive in vitro because it is released to bind to GroES . GroELSR2 has an additional mutation in Glu191-Gly, which restores activity by reducing the affinity for GroES. These two mutants form ring structures and would be suitable for use as frameworks.

Some naturally occurring framework molecules are bacteriophage products: for this reason, naturally occurring antibodies such as these frameworks are rare. This fact encourages the use of frame fusions as vaccine agents. Gp31 T4 with a suppressed loop has no biological activity (except as a dominant intracellular or negative vaccine against T4 bacteriophage) thus minimizing harmful effects on the host. However, the insertion of appropriate sequences encoding polypeptides can confer biological activity on the new proteins. In fact, biological activity can be improved by insertion into the shell protein.

The affinity of antibodies or antibody fragments for antigens (eg, human serum albumins) can be increased by oligomerization according to the present invention. The antibody fragments can be fragments such as Fv, Fab and F (ab ’) 2 fragments or any derivative thereof, such as single chain Fv fragments. Antibodies or antibody fragments may be non-recombinant, recombinant or humanized. The antibody can be of any immunoglobulin isotype, e.g. eg, IgG, IgM, and so on.

In some embodiments, the antibody fragments may be camelized VH domains. It is known that the main intermolecular interactions between antibodies and their similar antigens are mediated through CDR3 VH.

The use of GroEL and / or GroES framework molecules (cpn10) as described below and as is known in the art provides for oligomerization of VH domains, or CDR3 domains of VH, to produce a high affinity oligomer. Two or more domains may be included in said oligomer; in an oligomer based on a cpn10 framework, up to 7 domains can be included, to form a heptameric oligomeric molecule (heptabody) that is bound to a target protein (e.g., human serum albumin).

In order to impart and / or optimize the affinity of certain framework polypeptides / oligomers for a target protein

(e.g., human serum albumin), a variation can be introduced into heterologous polypeptides inserted into the framework polypeptides, so that the specificity and / or affinity of said polypeptides / oligomers for their ligands / substrates can be examined and / or mapped . Variants of the same loop can be produced, or a set of different standard loops can be devised, in order to quickly assess the affinity of a new polypeptide for a target protein (eg, human serum albumin). Variants can be produced by randomization of sequences according to known techniques, such as PCR. They can be subjected to selection by means of a screening protocol, such as phage display, before incorporation into protein frameworks.

An "oligomerizable framework," as referred to herein, is a polypeptide that is capable of oligomerizing or being oligomerized to form a framework and with which a heterologous polypeptide can be fused, preferably covalently, without suppressing oligomerization capabilities. . Thus, it provides a "framework" by means of which the polypeptides can be arranged in multimers according to the present invention. Optionally, parts of the natural polypeptide from which the framework was obtained can be removed, for example by substitution by the heterologous polypeptide that will be presented in the framework.

Monomers are polypeptides that possess the possibility of oligomerizing or being oligomerized. Oligomerization can be performed by incorporating, in the polypeptide, an oligomerizable shell subunit that will be oligomerized with additional shell subunits if combined with them. Optionally, oligomerization can be carried out through the use of linker groups recognized in the art for monomer binding.

As used herein, "oligomer" is synonymous with "polymer" or "multimer" and is used to indicate that the object in question is not monomeric. Thus, oligomeric polypeptides comprise at least two monomer units linked together by covalent or non-covalent means. The number of monomer units used will depend on the intended use of the oligomer, and may be between 2 and 20 or more. Optionally, it is between 5 and 10, and preferably is about 7.

Phage Presentation

Phage display technology has proven enormously useful in biological research. It allows to select ligands among large libraries of molecules. Frame technology can dominate phage display power in an exceptionally advantageous manner. The cpn10 molecules can be presented as monomers in bacteriophages fd, similar to the presentation of single-chain Fv molecules. Insertion libraries (instead of the highly mobile loop, eg, using CDR3 polypeptides obtained from human serum albumin binding antibodies) are constructed by standard methods, and the resulting libraries are screened for the molecules of interest. This selection is based on affinity. After identification of molecules that have affinity for the target protein (e.g., human serum albumin), potentially through one or more iterations of mutagenesis, expression (the GroEL, GroEL 57.5 kDa and GroES proteins of 10 kDa, can be expressed and purified as described above (Chatellier et al. 1998 Proc. Natl. Acad. Sci. USA 95: 9861-9866; Corrales and Fersht 1996 1: 265-273), or by any method recognized in the technique) and affinity screening, said molecules can undergo oligomerization, thus taking advantage of the avidity of said molecules. Optionally, some selected monomers may cross-link or oligomerize their binding participants.

Fibronectin

Example 26: Generation of a double specific ligand comprising a non-immunoglobulin framework of serum albumin binding fibronectin by means of CDR grafting

The CDR domains of dAb7h14 are used to construct a fibronectin non-immunoglobulin framework polypeptide that binds to human serum albumin in the following manner. The sequences of CDR1 (RASQWIGSQLS; SEQ ID NO.:__), CDR2 (WRSSLQS; SEQ ID NO.:__) and CDR3 (AQGAALPRT; SEQ ID NO.:__) of dAb7h14 are grafted into 10Fn3 replacing amino acid residues of natural 10Fn3 at positions 2131 (the BC loop), 51-56 (the DE loop) and 76-88 (the FG loop), respectively. Biacore real-time binding analysis is performed to assess whether human serum albumin is specifically bound to an immobilized fibronectin derived polypeptide comprising the human serum anti-albumin CDR domains of dAb7h14. (One skilled in the art will recognize that binding affinity can be assessed using any appropriate method, including, e.g., precipitation of labeled human serum albumin, competitive Biacore assay, etc.). If a low or no human serum albumin affinity is detected (e.g., Kd values in the range of M or higher), at least one of a set of strategies is used to improve serum albumin binding properties human CDR-grafted fibronectin polypeptide, which includes any of the following methods that contribute to binding affinity.

The length or lengths of dAb7h14 CDR-grafted regions of the fibronectin polypeptide corresponding to solvent-exposed loop regions in the natural fibronectin polypeptide are adjusted through the use of linker group polypeptides. For example, the CDR3 peptide sequence of nine CDR3 amino acid residues of dAb7h14 extends to 13 amino acid residues in length using amino acid linker groups of, e.g. eg, zero to four length residues located on one and / or the two flanks of the N or C ends of the CDR3 polypeptide sequence dAb7h14, thus achieving a total grafted peptide sequence length of 13 amino acids in the CDR3 grafted domain corresponding to the FG loop in the natural fibronectin sequence. Said use of polypeptide or linker group polypeptides is optionally combined with mutagenesis of linker group sequences, CDR sequences and / or non-CDR fibronectin sequences (for example, using mutagenic optimization procedures as described below), in order to improve the capacity of human serum albumin binding of CDR-grafted fibronectin polypeptides (e.g., by optimization of CDR and fibronectin sequences in the CDR-grafted fibronectin polypeptides). The polypeptide linker groups used for this purpose have a predetermined sequence, or, optionally, are selected from a population of polypeptide sequences of randomized linker groups by evaluating the human serum albumin binding capabilities of CDR-grafted fibronectin polypeptides containing linker group Optimization methods are performed in parallel and / or iteratively. Both parallel and iterative optimization processes (e.g., affinity maturation) employ selection methods as described below and / or as known in the art as useful for optimizing the binding properties of polypeptides. .

Human serum albumin binding of the CDR-grafted fibronectin polypeptide or polypeptides presenting dAb7h14 CDR is optimized by mutagenesis, optionally in combination with parallel and / or iterative selection methods as described below and / or known by others. means in the art. The 10Fn3 framework polypeptide domains surrounding the grafted CDR dAb7h14 polypeptide sequences are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed in the 10Fn3 polypeptide sequence at amino acids 1-9, 44-50, 61-54, 82-94 (beta sheet edges); 19, 21, 30-46 (even), 79-65 (odd) (faces accessible to the solvents of the two beta sheets); and 14-16 and 36-45 (bonds accessible to non-CDR type solvents and beta spins), and optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. Optionally, dAb7h14 CDR polypeptide domains presented in the CDR grafted fibronectin polypeptide are subjected to mutagenesis by, e.g. eg, random mutagenesis, NNK mutagenesis, review mutagenesis and / or other method recognized by the art. To perform said mutagenesis methods, PCR is optionally used, which results in the generation of a variety of sequences through targeted sequences in the CDR-grafted fibronectin polypeptides. These strategies are similar to those described below for generating dAb libraries. In addition to methods of random and / or review mutagenesis, directed mutagenesis of targeted amino acid residues is used where the structural information establishes specific amino acid residues that are essential for human serum albumin binding.

Fibronectin polypeptides comprising grafted dAb7h14 CDR sequences designed as described above are subjected to parallel and / or iterative selection methods to identify those fibronectin polypeptides that are optimized for human serum albumin binding. For example, after the production of a library of fibronectin polypeptide sequences grafted with CDR dAb7h14, this library of said polypeptides is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized fibronectin derived polypeptides that ideally reach Kd values of serum albumin binding affinity in the range of nM or better. After identification of fibronectin-derived polypeptides that bind human serum albumin, said polypeptides are used to generate specific double ligand compositions by any of the methods described below.

Example 27: Generation of a double specific ligand comprising a non-immunoglobulin framework of serum albumin binding fibronectin by means of selecting serum albumin binding moieties

The natural fibronectin protein, specifically the 10Fn3 fibronectin polypeptide, is subjected to library selection techniques and, optionally, affinity maturation in order to produce non-immunoglobulin framework molecules of serum albumin-binding fibronectin for use in the specific double ligands described herein.

A real-time binding analysis is performed by Biacore to assess whether human serum albumin is specifically bound to fibronectin polypeptide and / or immobilized natural fibronectin derivative. After detection of null or low binding affinity (e.g., Kd values in the range of M or higher) of a human serum albumin fibronectin polypeptide, at least one of a set of strategies is used to imparting human serum albumin binding properties to the fibronectin polypeptide, which includes one or more of the following methods that contribute to binding affinity.

Human serum albumin binding of polypeptide or fibronectin binding polypeptides is achieved and optimized by mutagenic methods, optionally in combination with parallel and / or iterative selection methods as described below and / or known by other means in The technique. The 10Fn3 framework polypeptide domains are subjected to randomized and / or NNK mutagenesis, performed as described below. Said mutagenesis is performed on the entire 10Fn3 polypeptide or on specific sequences in the 10Fn3 polypeptide, which includes amino acids 1-9, 44-50, 61-54, 82-94 (beta sheet edges); 19, 21, 30-46 (even), 79-65 (odd) (solvent-accessible faces of the two beta sheets); 21-31, 51-56, 76-88 (solvent-accessible bonds of the CDR type); and 14-16 and 36-45 (other solvent-accessible bonds and beta spins), and optionally randomized in order to obtain new or improved polypeptides binding to human serum albumin. PCR is optionally used to perform said mutagenesis methods, which results in the generation of a variety of sequences through sequences directed at the fibronectin polypeptides. (These strategies are similar to those described below for generating dAb libraries). In addition to random methods of mutagenesis, directed mutagenesis of targeted amino acid residues is employed where the structural information establishes specific amino acid residues of fibronectin polypeptides that are critical for the binding of human serum albumin.

Fibronectin polypeptides designed as described above are subjected to parallel and / or iterative selection methods to identify those fibronectin polypeptides that are optimized for binding to human serum albumin. For example, after the production of a library of mutagenized fibronectin polypeptide sequences, said polypeptide library is presented in phages and subjected to multiple selection batches that require serum albumin binding and / or proliferation, as described below for selection of serum albumin binding dAb from dAb libraries. Optionally, the selection is made against serum albumin immobilized in immunotubes or against biotinylated serum albumin in solution. Optionally, binding affinity is determined using surface plasmon resonance (SPR) and Biacore (Karlsson et al., 1991), using a Biacore system (Uppsala, Sweden), with fully optimized fibronectin derived polypeptides that ideally reach Kd values of serum albumin binding affinity in the range of nM or better.

After identification of fibronectin polypeptides that bind to human serum albumin, said polypeptides are used to generate specific double ligand compositions by any of the methods described below.

Fibronectin non-immunoglobulin frameworks

A non-immunoglobulin framework comprising fibronectin, or a functional moiety and / or fragment thereof, can be designed to bind to serum albumin. A non-immunoglobulin framework structure obtained from the type III fibronectin module (Fn3) is used. The type III fibronectin module is a common domain in structural and mammalian blood proteins, which appears more than 400 times in the protein sequence database and, as estimated, appears in 2% of all the proteins sequenced to date. Proteins that include an Fn3 module sequence include fibronectins, tenascin, intracellular cytoskeletal proteins and prokaryotic enzymes (Bork and Doolittle, Proc. Natl. Acad. Sci. USA 89: 8990, 1992; Bork et al., Nature Biotech. 15: 553, 1997; Meinke et al., J. Bacteriol. 175: 1910, 1993; Watanabe et al., J. Biol. Chem. 265: 15659, 1990). A framework of fibronectin non-immunoglobulins in particular is the tenth module of human Fn3 (10Fn3), which comprises 94 amino acid residues. The total folding of this domain is closely related to that of the smallest functional antibody fragment, the heavy chain variable region, which comprises the complete antigen recognition unit in camel and llama IgG. The main differences between the camel and llama domains and the 10Fn3 domain are that (i) 10Fn3 has fewer beta chains (seven versus. Nine) and (ii) the two beta sheets packed together are connected by a disulfide bridge in the camel and llama domains, but not in 10Fn3.

The three 10Fn3 bonds that correspond to the IgG heavy chain antigen binding bonds are located between amino acid residues 21-31 (BC), 51-56 (DE) and 76-88 (FG) (see Figure 3 of US Patent No. 7,115,396, the complete content of which is incorporated herein by reference). The lengths of the BC and DE bonds, 11 and 6 residues, respectively, are within the narrow range of the corresponding antigen recognition loops that are present in the heavy antibody chains, i.e., 7-10 and 4-8. waste, respectively. Therefore, a CDR grafting strategy can easily be used to introduce heavy chain CDR sequences into these domains. Additionally and / or alternatively, these two loops can be subjected to the introduction of genetic variability by any method recognized in the art (eg, site-directed mutagenesis, revision or other method, randomization, etc.). and, optionally, the resulting polypeptide can be screened for high antigen affinity. (Alternatively, the introduction of genetic variability and / or selection procedures can be used to identify compositions with lower binding affinity and / or optimized properties such as stability, toxicity, etc.). Through the use of these methods, the BC and DE fibronectin bonds can be designed to form contacts with antigens equivalent to the contacts of the corresponding CDR1 and CDR2 domains in the antibodies.

Unlike the BC and DE loops, the 10Fn3 FG loop is 12 residues in length, while the corresponding loop in the antibody heavy chains is between 4-28 residues. Accordingly, to optimize antigen binding, the length of the FG loop of 10Fn3 can be varied (e.g., by the use of randomization and / or the use of linker group polypeptide sequences (which can also be randomized)) as well as in the sequence to cover the CDR3 length range of 4-28 residues to obtain as much flexibility and affinity as possible in antigen binding. In fact, for those methods in which the CDRs are grafted directly into a fibronectin shell and in which a natural fibronectin shell is selected and / or optimized for serum albumin binding (or other target antigen), the lengths thus as the sequences of the CDR type loops of the antibody emulation can be randomized during affinity maturation in vitro or in vivo (as described in more detail below).

The tenth type III domain of human fibronectin, 10Fn3, quickly folds back even at low temperature; its main conformation has recovered in 1 second at 5 ° C. The thermodynamic stability of 10Fn3 is high (ΔGu = 24 kJ / mol = 5.7 kcal / mol), a fact correlated with its high melting temperature of 110 ° C.

One of the physiological functions of 10Fn3 is as a subunit of fibronectin, a glycoprotein that exists in a soluble form in body fluids and in an insoluble form in the extracellular matrix (Dickinson et al., J. Mol. Biol. 236: 1079 , 1994). A 220-250 Kd fibronectin monomer contains 12 type I modules, two type II modules and 17 type III fibronectin modules (Potts and Campbell, Curr. Opin. Cell Biol. 6: 648, 1994). In the binding of fibronectin to integrins, heparin and chondroitin sulfate intervene different type III modules. It was determined that 10Fn3 acts as a mediator in cell adhesion through an Arg-Gly-Asp (RGD) motif of integrin binding in one of its exposed bonds. Other similar RGD motifs have been shown to be involved in integrin binding by other proteins, such as fibrinogen, von Willebrand factor and vitronectin (Hynes et al., Cell 69:11, 1992). No other matrix or cell binding function has been described for 10Fn3.

The observation that 10Fn3 has only slightly more adhesive activity than a short peptide containing RGD is consistent with the conclusion that the 10Fn3 cell binding activity is located in the RGD peptide rather than distributed throughout the 10Fn3 structure ( Baron et al., Biochemistry 31: 2068, 1992). The fact that it is unlikely that 10Fn3 without the RGD motif binds to other plasma proteins or the extracellular matrix makes 10Fn3 a useful framework for replenishing antibodies. In addition, the presence of 10Fn3 in the natural fibrinogen in the bloodstream indicates that it is likely that 10Fn3 is immunogenic in the organism of origin.

In addition, it has been shown that the 10Fn3 framework has randomly tolerated exposed loop sequences, which facilitates the generation of various antibody emulation reserves. This determination was made by examining the flexibility of the 10Fn3 sequence. In particular, the human 10Fn3 sequence was aligned with the fibronectin sequences from other sources as well as related protein sequences, and a correspondence was established between the results of this alignment and the three-dimensional structure of the human 10Fn3 domain. This alignment revealed that most of the conserved residues were present in the core of the interleaving of the beta sheets, while the highly variable residues were located at the edges of the beta sheets, which include the N and C ends, on the faces solvent-accessible from the two beta sheets, and in three solvent-accessible bonds that acted as hypervariable bonds for affinity maturation of antibody emulators. In view of these results, it was determined that the randomization of these three loops was unlikely to have an adverse effect on the folding or overall stability of the 10Fn3 framework itself.

For the human 10Fn3 sequence, this analysis indicated that at least amino acids 1-9, 44-50, 61-54, 82-94 (beta sheet edges); 19, 21, 30-46 (even), 79-65 (odd) (solvent-accessible faces of the two beta sheets); 21-31, 51-56, 76-88 (solvent-accessible bonds of the CDR type); and 14-16 and 36-45 (other solvent-accessible bonds and beta spins) could be randomized to obtain new compound binding proteins by evolution

or improved. In addition, as discussed above, alterations in the lengths of one or more bonds exposed to solvents could also be included in such directed evolution methods.

Alternatively, changes in β-leaf sequences could also be used to obtain new proteins by evolution. These mutations modify the framework and thus indirectly alter the loop structure (s). If this approach is adopted, the mutations should not saturate the sequence, but rather few mutations would be introduced. Preferably, no more than 3-20 changes in the sequences of the β sheets should be introduced by this approach.

A sequence variation can be introduced by any technique that includes, for example, Taq polymerase mutagenesis (Tindall and Kunkel, Biochemistry 27: 6008 (1988)), fragment recombination or a combination thereof. Similarly, an increase in the structural diversity of libraries can be used, for example, by modifying the length as well as the sequence of the CDR and / or CDR type presentation loops, or by structural redesign based on the advantageous framework mutations present in selected reserves, in order to introduce additional improvements in non-immunoglobulin frameworks.

Fusion proteins comprising fibronectin shell polypeptides

The fibronectin sheath polypeptides described herein can be fused with other protein domains. For example, the fibronectin sheath polypeptides identified for human serum albumin binding can be fused with individual domains of variable heavy chains, or an antigen binding fragment thereof, in order to generate a specific double ligand comprising a serum albumin binding residue based on fibronectin. Fibronectin framework polypeptides can be further integrated with the human immune response by fusing the constant region of an IgG (Fc) with a fibronectin framework polypeptide, such as a 10Fn3 module, preferably through the C-terminus of 10Fn3. The Fc of said 10Fn3-Fc fusion molecule activates the complement component of the autoimmune response and can serve to increase the therapeutic value of the designed fibronectin polypeptide. Similarly, a fusion between a fibronectin framework polypeptide, such as 10Fn3, and a complement protein, such as C1q, for target cells, can be used, and a fusion between a fibronectin framework polypeptide, such as 10Fn3, and a toxin can be used to destroy specifically the cells that carry a particular antigen. Any of these fusions can be generated by standard techniques, for example, by expression of the fusion protein from a recombinant fusion gene constructed using publicly available gene sequences and / or as described below by other means.

Frame Multimers

In addition to monomers, any of the fibronectin framework constructs described herein can be generated as dimers or multimers of frameworks as a means of increasing valence and thus the avidity of antigen binding (e.g., serum albumin) . Such multimers can be generated through covalent bonding. For example, the individual 10Fn3 modules can be joined by mimicking natural binding at the C-terminus to N 8Fn3-9Fn3-10Fn3 or by mimicking the antibody dimers that are held together through their constant regions. A 10Fn3-Fc construction can be used to design dimers of the general scheme of 10Fn3Fc :: Fc-10Fn3. The links designed in the Fc :: Fc interface can be covalent or non-covalent. In addition, dimerization or multimerization participants other than Fc can be used, such as other non-immunoglobulin framework residues and / or immunoglobulin-based antigen binding residues, in hybrids, such as 10Fn3 hybrids, to create such higher order structures. Other examples of multimers include the individual variable domains described herein.

In particular examples, covalently bonded multimers can be generated by constructing fusion genes encoding the multimer or, alternatively, designing codons for cysteine residues in monomeric sequences and allowing the formation of disulfide bonds between expression products. Non-covalently bound multimers can also be generated by various techniques. They include the introduction, in monomeric sequences, of codons that correspond to residues with positive and / or negative charge and that allow interactions between these residues to occur in the expression products (and therefore between the monomers). This approach can be simplified by taking advantage of the charged residues naturally present in a monomer subunit, for example, negatively charged fibronectin residues. Another means of generating compositions with non-covalent bonds comprising fibronectin framework polypeptides is to introduce, in the monomer gene (for example, at the amino or carboxy ends), the coding sequences for proteins or protein domains with which they are He knows they interact. Such proteins or protein domains include spiral-spiral motifs, leucine zipper motifs and any of the numerous protein subunits (or fragments thereof) for which it is known to direct the formation of higher order dimers or multimers.

Fibronectin type molecules

Although 10Fn3 represents a preferred framework for the generation of antibody emulators, other molecules can be substituted for 10Fn3 in the molecules described herein. These include, without limitation, 1Fn3-9Fn3 and 11Fn3-17Fn3 human fibronectin modules as well as related Fn3 modules from nonhuman and prokaryotic animals. In addition, the Fn3 modules of other proteins with sequence homology with 10Fn3, such as tenascins and undulins, can also be used. Other example frameworks that have immunoglobulin-like folds (but with sequences that are not related to the VH domain) include N-cadherin, ICAM-2, titin, GCSF receptor, cytokine receptor, glucosidase inhibitor, E-cadherin and Antibiotic chromoprotein Additional domains with related structures of myelin membrane adhesion molecule P0, CD8, CD4, CD2, MHC class I, T-cell antigen receptor, CD1, C2 and I-set domains of VCAM-1, immunoglobulin domain can be obtained I-set of myosin-binding protein C, immunoglobulin domain I-set of myosin-binding protein H, immunoglobulin domain I-set of telocin, telicin, NCAM, twitchin, neurogliano, growth hormone receptor, receptor Erythropoietin, prolactin receptor, GC-SF receptor, interferon-gamma receptor, β-galactosidase / glucuronidase, β-glucuronidase and transglutaminase. Alternatively, any other protein that includes one or more immunoglobulin type folds can be used. Such proteins can be identified, for example, using the SCOP program (Murzin et al.,

J. Mol. Biol. 247: 536 (1995); Lo Conte et al., Nucleid Acid Res. 25: 257 (2000).

In general, any molecule that shows a structural relationship with the VH domain (as identified, for example, using the above SCOP software) can be used as a non-immunoglobulin framework. These molecules, such as fibronectin, can include three bonds at the N-terminus of the molecule and three bonds at the C-terminus, each of which can be randomized to create various libraries; alternatively, larger domains can be used, which have a greater number of loops, as long as there are a number of said randomizable surface loops located close enough in space so that it can participate in antigen binding. Examples of polypeptides that possess more than three loops arranged close to each other include receptor T-cell antigens and superoxide dismutase, each having four loops that can be randomized; and an Fn3 dimer, tissue factor domains and cytokine receptor domains, each of which has three sets of two similar domains in which three randomizable bonds are part of the two domains (to bring the total number of ties to six ).

In yet another alternative, any protein having variable loops placed sufficiently close in space for the production of candidate binding proteins can be used. For example, large proteins can be used that have links accessible to spatially related solvents, even if they are not structurally related to an immunoglobulin folding. Exemplary proteins include, without limitation, cytochrome F, green fluorescent protein, GroEL and thaumatine. The bonds presented by these proteins may be randomized and higher binders are selected from a randomized library as described herein, e.g. ex. in Example 1. Due to their size, molecules showing an antigen-binding surface considerably larger than that found in an antibody-antigen interaction can be obtained. Other useful frameworks of this type can also be identified using the SCOP program (Murzin et al., J. Mol. Biol. 247: 536 (1995)) to search among candidate proteins that have numerous loops, especially loops placed between parallel beta sheets. or a series of alpha helices.

The modules of different organisms and progenitor proteins may be the most appropriate for different applications. For example, in the design of a fibronectin shell polypeptide, they may be the most convenient for generating that protein from a fibronectin molecule or of the natural fibronectin type for the organism for which the therapeutic action is intended. In contrast, the organism of origin is less important or even irrelevant for fibronectin frameworks that will be used for in vitro applications, such as diagnosis,

or as research reagents.

For any of these molecules, libraries can be generated and used to select binding proteins by any of the methods described herein.

Targeted evolution of framework-based binding proteins

The non-immunoglobulin frameworks described herein can be used in any technique to evolve new or improved binding proteins. In a particular example, the binding target (for example, serum albumin) is immobilized on a solid support, such as a column resin or a microtiter plate well, and the target is contacted with a protein library of binding based on candidate non-immunoglobulin framework. Said library may consist of fibronectin framework clones, such as 10Fn3 clones constructed from natural (wild type) 10Fn3 framework through randomization of the sequence and / or length of CDR 10Fn3 type loops. If desired, this library may be an RNA protein fusion library generated, for example, by the techniques described in Szostak et al., US document. serial number 09 / 007.005 and document serial number 09 / 247.190; Szostak et al., WO98 / 31700; and Roberts & Szostak, Proc. Natl Acad. Sci. USA (1997) vol. 94, p. 12297-12302. Alternatively, it can be a DNA protein library (for example, as described in Lohse, DNA-Protein Fusions and Uses Thereof, US document No. 60 / 110,549, US document No. 09 / series 459,190, and WO-00/32823). The fusion library is incubated with the immobilized target, the support is washed to remove non-specific binders and the firmer binders are eluted under very severe conditions and subjected to PCR to retrieve the sequence information or to create a new library of binders. which can be used to repeat the selection process, with or without additional sequence mutagenesis. A series of selection batches may be performed until binders of sufficient affinity for the antigen (eg, serum albumin) are obtained.

In a particular example, the 10Fn3 framework can be used as the selection target. For example, if a protein that binds to a specific peptide sequence (e.g., serum albumin) presented in a loop of ten residues is required, a single 10Fn3 clone is constructed in which one of its bonds has been set a length of ten and for the desired sequence. The new clone is expressed in vivo and purified, and then immobilized on a solid support. An RNA protein fusion library based on an appropriate framework is then allowed to interact with the support, which is then washed, and the desired molecules eluted and re-selected as described above.

Similarly, the frameworks described herein, for example, the 10Fn3 framework, can be used to find natural proteins that interact with the peptide sequence shown by the framework, for example, in a 10Fn3 loop. The framework protein, such as the 10Fn3 protein, is immobilized as described above, and an RNA protein fusion library for binders in the presented loop is screened. The binders are enriched through multiple batches of selection and identified by DNA sequencing.

In addition, in the above approaches, although RNA protein libraries represent example libraries for directed evolution, any type of framework-based library can be used in the selection methods described herein.

Use of fibronectin framework polypeptides

The fibronectin sheath polypeptides described herein can be evolved to bind to serum albumin or to any antigen of interest. Said fibronectin shell proteins have thermodynamic properties superior to those of natural antibodies and can be rapidly evolved in vitro. Consequently, these fibronectin framework polypeptides can be used to produce binding domains for use in the fields of research, therapy and diagnosis.

Mutagenic affinity maturation

The selections described herein can also be combined with mutagenesis after all or a subset of the selection steps to further increase the diversity of libraries. Affinity maturation methods may employ, e.g. eg, error-prone PCR (Cadwell and Joyce, PCR Methods Appl 2:28 (1992)) or an alternative form of random mutagenesis, NNK mutagenesis as described below, review mutagenesis (where framework polypeptides are designed for CDR graft fibronectin to optimize antigen binding through the use of naturally occurring CDR diversity; see, e.g., WO-06/023144, incorporated herein by reference) and / or others Mutagenesis approaches recognized in the art to create diversity of polypeptides, which is combined with one or more selection batches for antigen binding affinity.

Any of the framework proteins described below can be combined with another for use, e.g. eg, in the specific double ligand compositions described herein. For example, CDR can be grafted onto a CTLA-4 framework and used together with VH or VL antibody domains to form a multivalent ligand. Similarly, fibronectin, lipocalin, affibody molecules and other frameworks can be combined.

Appendix 1; polypeptides that enhance half-life in vivo.

Alpha-1 Glycoprotein (Orosomucoid) (AAG) Alpha-1 Antiquiromotripsin (ACT) Alpha-1 Antitrypsin (AAT) Alpha-1 Microglobulin (HC Protein) (AIM) Alpha-2 Macroglobulin (A2M) Antithrombin III (AT III) Apolipoprotein A-1 (Apo A-1) Apoliprotein B (Apo B) Beta-2-microglobulin (B2M) Ceruloplasmin (Cp) Complement component (C3) Complement component (C4) C1 esterase inhibitor (C1 INH) C-reactive protein (PCR) Cystatin C (Cys C) Ferritin (FER) Fibrinogen (FIB) Fibronectin (FN) Haptoglobin (Hp) Hemopexin (HPX)

Immunoglobulin A (IgA) Immunoglobulin D (IgD) Immunoglobulin E (IgE) Immunoglobulin G (IgG)

5 Immunoglobulin M (IgM) Light chains of immunoglobulins (kappa / lambda) Lipoprotein (a) [Lp (a)] Mannose binding protein (MBP) Myoglobin (Myo)

10 Plasminogen (PSM) Prealbumin (Transthyrethrin) (PAL) Retinol-binding protein (RBP) Rheumatoid factor (RF) Serum amyloid A (SAA)

15 Soluble transferrin receptor (sTfR) Transferrin (Tf)

Appendix 2

Pairing

References of therapeutic interest.

TNF ALPHA / TGF-β

• TGF-b and TNF when injected into the ankle joint of the collagen-induced arthritis model significantly improved joint inflammation. In mice without a collagen challenge test, no effect occurred.

TNF ALPHA / IL-1

• TNF and IL-1 act synergistically in the condition of uveitis. • TNF and IL-1 act synergistically in malaria disease (hypoglycemia, NO). • TNF and IL-1 act synergistically in the induction of polymorphonuclear cell migration (PMN) in case of inflammation. • IL-1 and TNF act synergistically to induce PMN infiltration in the peritoneum. • IL-1 and TNF act synergistically to induce the secretion of IL-1 by endothelial cells. Important in inflammation. • IL-1 or TNF alone induced some cellular infiltration into the patellar synovial fluid. IL-1 induced PMN, TNF-monocytes. Together they induced more severe inflammation due to the increase in PMN. • The circulating myocardial depressant substance (present in sepsis) has low levels of IL-1 and TNF that act synergistically.

TNF ALPHA / IL-2

• Mostly related to the synergistic activation of cytolytic T lymphocytes.

TNF ALPHA / IL-3

• Synergy of interleukin 3 and tumor necrosis factor alpha in stimulating clonal growth of acute myelogenous leukemia blasts is the result of induction of secondary hematopoietic cytokines by tumor necrosis factor alpha. • Cancer Res. 1992 Apr 15; 52 (8): 2197-201.

TNF ALPHA / IL-4

• IL-4 and TNF act synergistically to induce VCAM expression in endothelial cells. It is derived that they have a function in asthma. The same is true for synovial fluid, which is involved in RA. • TNF and IL-4 act synergistically to induce expression of IL-6 in keratinocytes. • Sustained elevated levels of VCAM-1 in cultured fibroblast-like synoviocytes can be achieved by TNF-alpha in combination with either IL-4 or IL-13 through increased mRNA stability. Am J Pathol. 1999 Apr; 154 (4): 1149-58

TNF ALPHA / IL-5

• Relationship between the tumor necrosis factor system and the serum interleukin4, interleukin-5, interleukin-8, eosinophil cationic protein, and immunoglobulin E levels in the bronchial hyperreactivity of adults and their children. Allergy Asthma Proc. 2003 Mar-Apr; 24 (2): 111-8.

TNF ALPHA / IL-6

• TNF and IL-6 are potent growth factors for OH-2, a novel human myeloma cell line. Eur J Haematol. 1994 Jul; 53 (1): 31-7.

TNF ALPHA / IL-8

• TNF and IL-8 act synergistically with PMN to activate the platelets involved in acute respiratory distress syndrome. • See IL-5 / TNF (asthma). Synergism between interleukin-8 and tumor necrosis factor-alpha for neutrophil-mediated platelet activation. Eur Cytokine Netw. 1994 Sep-Oct; 5 (5): 455-60. (adult respiratory distress syndrome (ARDS))

TNF ALPHA / IL-9

TNF ALPHA / IL-10

• IL-10 induces and acts synergistically with TNF in inducing HIV expression in chronically infected T lymphocytes.

TNF ALPHA / IL-11

• Cytokines synergistically induces osteoclast differentiation: support by immortalized or normal calvarial cells. Am J Physiol Cell Physiol. 2002 Sep; 283 (3): C679-87. (Bone loss)

TNF ALPHA / IL-12

TNF ALPHA / IL-13

• Sustained elevated levels of VCAM-1 in cultured fibroblast-like synoviocytes can be achieved by TNF-alpha in combination with either IL-4 or IL-13 through increased mRNA stability. Am J Pathol. 1999 Apr; 154 (4): 1149-58. • Interleukin-13 and tumor necrosis factor-alpha synergistically induces eotaxin production in human nasal fibroblasts. Clin Exp Allergy. 2000 Mar; 30 (3): 348-55. • Interleukin-13 and tumor necrosis factor-alpha synergistically induces eotaxin production in human nasal fibroblasts. Clin Exp Allergy. 2000 Mar; 30 (3): 348-55 (allergic inflammation) • Implications of serum TNF-beta and IL-13 in the treatment response of childhood nephrotic syndrome. Cytokine 2003 Feb 7; 21 (3): 155-9.

TNF ALPHA / IL-14

• Effects of inhaled tumor necrosis factor alpha in subjects with mild asthma. Thorax 2002 Sep; 57 (9): 774-8.

TNF ALPHA / IL-15

• Effects of inhaled tumor necrosis factor alpha in subjects with mild asthma. Thorax 2002 Sep; 57 (9): 774-8.

TNF ALPHA / IL-16

• Tumor necrosis factor-alpha-induced synthesis of interleukin-16 in airway epithelial cells: priming for serotonin stimulation. Am J Respir Cell Mol Biol. 2003 Mar; 28 (3): 354-62. (respiratory tract inflammation) • Correlation of circulating interleukin 16 with proinflammatory cytokines in patients with rheumatoid arthritis. Rheumatology (Oxford). 2001 Apr; 40 (4): 474-5. There is no summary available. • Interleukin 16 is up-regulated in Crohn’s disease and participates in TNBS colitis in mice. Gastroenterology 2000 Oct; 119 (4): 972-82.

TNF ALPHA / IL-17

• Inhibition of interleukin-17 prevents the development of arthritis in vaccinated mice challenged with Borrelia burgdorferi. Infect Immun. 2003 Jun; 71 (6): 3437-42. • Interleukin 17 synergises with tumor necrosis factor alpha to induce cartilage destruction in vitro. Ann Rheum Dis. 2002 Oct; 61 (10): 870-6. • A role of GM-CSF in the accumulation of neutrophils in the airways caused by IL17 and TNF-alpha. Eur Respir J. 2003 Mar; 21 (3): 387-93. (Inflammation of the respiratory tract) • Summary Interleukin-1, tumor necrosis factor alpha, and interleukin-17 synergistically upregulate nitric oxide and prostaglandin E2 production in explants of human osteoarthritic knee menisci. Arthritis Rheum 2001 Sep; 44 (9): 2078-83.

TNF ALPHA / IL-18

• Association of interleukin-18 expression with enhanced levels of both interleukin-1beta and tumor necrosis factor alpha in knee synovial tissue of patients with rheumatoid arthritis. Arthritis Rheum 2003 Feb; 48 (2): 339-47. • Summary Elevated levels of interleukin-18 and tumor necrosis factor-alpha in serum of patients with type 2 diabetes mellitus: relationship with diabetic nephropathy. Metabolism 2003 May; 52 (5): 605-8.

TNF ALPHA / IL-19

• Summary IL-19 induces production of IL-6 and TNF-alpha and results in cell apoptosis through TNF-alpha. J Immunol. 2002 Oct 15; 169 (8): 4288-97.

TNF ALPHA / IL-20

• Cytokines Summary: IL-20 -a new effector in skin inflammation. Curr Biol. 2001 Jul 10; 11 (13): R531-4

TNF ALPHA / Complement

• Inflammation and coagulation: implications for the septic patient. Clin Infect Dis. 2003 May 15; 36 (10): 1259-65. Epub 2003 May 08. Review.

TNF ALPHA / IFN-γ

• MHC induction in the brain. • Synergy in antiviral response / induction of IFN-β • Neutrophil activation / respiratory impulse. • Activation of endothelial cells • Toxicities are observed when patients are treated with TNF / IFN-γ as antiviral therapy • Fractalcin expression by human astrocytes. • Many articles on inflammatory responses, p. eg, LPS, also macrophage activation. • Anti-TNF and anti-IFN-γ act synergistically to protect mice from deadly endotoxemia.

TGF-β / IL-1

• Synthesis of prostaglandins by osteoblasts • Production of IL-6 by intestinal epithelial cells (inflammation model) • Stimulates IL-11 and IL-6 in lung fibroblasts (inflammation model) • Production of IL-6 e IL-8 in the retina

TGF-β / IL-6

• Proliferation of chondrosarcoma

IL-1 / IL-2

• B lymphocyte activation • LAK lymphocyte activation • T lymphocyte activation • IL-1 synergy with IL-2 in the generation of lymphokine activated killer cells is mediated by TNF-alpha and beta (lymphotoxin). Cytokine 1992 Nov; 4 (6): 479-87.

IL-1 / IL-3

IL-1 / IL-4

• Activation of B lymphocytes • IL-4 induces the expression of IL-1 in activation of endothelial cells.

IL-1 / IL-5

IL-1 / IL-6

• Activation of B lymphocytes • Activation of T lymphocytes (can replace accessory cells) • IL-1 induces the expression of IL-6 • Expression of serum amyloid and C3 (acute phase response) • HIV expression • Collagen decomposition of the cartilage.

IL-1 / IL-7

• IL-7 is requisite for IL-1-induced thymocyte proliferation. Involvement of IL-7 in the synergistic effects of granulocyte-macrophage colony-stimulating factor or tumor necrosis factor with IL-1. J Immunol. 1992 Jan 1; 148 (1): 99-105.

IL-1 / IL-8

IL-1 / IL-10

IL-1 / IL-11

• Cytokines synergistically induces osteoclast differentiation: support by immortalized or normal calvarial cells. Am J Physiol Cell Physiol. 2002 Sep; 283 (3): C679-87. (Bone loss)

IL-1 / IL-16

• Correlation of circulating interleukin 16 with proinflammatory cytokines in patients with rheumatoid arthritis. Rheumatology (Oxford). 2001 Apr; 40 (4): 474-5. No summary available.

IL-1 / IL-17

• Inhibition of interleukin-17 prevents the development of arthritis in vaccinated mice challenged with Borrelia burgdorferi. Infect Immun. 2003 Jun; 71 (6): 3437-42. • Contribution of interleukin 17 to human cartilage degradation and synovial inflammation in osteoarthritis. Osteoarthritis Cartilage. 2002 Oct; 10 (10): 799-807. • Summary Interleukin-1, tumor necrosis factor alpha, and interleukin-17 synergistically upregulate nitric oxide and prostaglandin E2 production in explants of human osteoarthritic knee menisci. Arthritis Rheum 2001 Sep; 44 (9): 2078-83.

IL-1 / IL-18

• Association of interleukin-18 expression with enhanced levels of both interleukin1beta and tumor necrosis factor alpha in knee synovial tissue of patients with rheumatoid arthritis. Arthritis Rheum 2003 Feb; 48 (2): 339-47.

IL-1 / IFN-g

IL-2 / IL-3

• Proliferation of T lymphocytes • Proliferation of B lymphocytes

IL-2 / IL-4

• Proliferation of B lymphocytes • Proliferation of T lymphocytes • (selectively inducing activation of CD8 and NK lymphocytes) IL-2R beta agonist P130 acts in synergy with IL-2, IL-4, IL-9, and IL-15: biological and molecular effects J Immunol. 2000 Oct 15; 165 (8): 4312-8.

IL-2 / IL-5

• B lymphocyte proliferation / Ig secretion • IL-5 induces IL-2 receptors in B lymphocytes

IL-2 / IL-6

• Development of cytotoxic T lymphocytes

IL-2 / IL-7

IL-2 / IL-9

• See IL-2 / IL-4 (NK lymphocytes)

IL-2 / IL-10

• B lymphocyte activation

IL-2 / IL-12

• IL-12 synergizes with IL-2 to induce lymphokine-activated cytotoxicity and perforin and granzyme gene expression in fresh human NK cells. Immunol cell. 1995 Oct 1; 165 (1): 33-43. (T lymphocyte activation)

IL-2 / IL-15

• See IL-2 / IL-4 (NK lymphocytes)

• (T cell activation and proliferation) IL-15 and IL-2: a matter of life and death for T cells in vivo. Nat Med. 2001 Jan; 7 (1): 114-8.

IL-2 / IL-16

• Synergistic activation of CD4 + T cells by IL-16 and IL-2. J Immunol. 1998 Mar 1; 160 (5): 2115-20.

IL-2 / IL-17

• Evidence for the early involvement of interleukin 17 in human and experimental renal allograft rejection. J Pathol. 2002 Jul; 197 (3): 322-32.

IL-2 / IL-18

• Interleukin 18 (IL-18) in synergy with IL-2 induces lethal lung injury in mice: a potential role for cytokines, chemokines, and natural killer cells in the pathogenesis of interstitial pneumonia. Blood. 2002 Feb 15; 99 (4): 1289-98.

IL-2 / TGF-β

• Control of CD4 effector fate: transforming growth factor beta 1 and interleukin 2 synergize to prevent apoptosis and promote effector expansion. J Exp Med. 1995 Sep 1; 182 (3): 699-709.

IL-2 / IFN-γ

• Ig secretion by B lymphocytes • IL-2 induces IFN-γ expression by T lymphocytes

IL-2 / IFN-α / β

• None

IL-3 / IL-4

• Act synergistically on mast cell growth • Synergistic effects of IL-4 and either GM-CSF or IL-3 on the induction of CD23 expression by human monocytes: regulatory effects of IFN-alpha and IFN-gamma. Cytokine 1994 Jul; 6 (4): 407-13.

IL-3 / IL-5

IL-3 / IL-6

IL-3 / IFN-γ

• IL-4 and IFN-gamma synergistically increase total polymeric IgA receptor levels in human intestinal epithelial cells. Role of protein tyrosine kinases. J Immunol. 1996 Jun 15; 156 (12): 4807-14.

IL-3 / GM-CSF

• Differential regulation of human eosinophil IL-3, IL-5, and GM-CSF receptor alphachain expression by cytokines: IL-3, IL-5, and GM-CSF down-regulate IL-5 receptor alpha expression with loss of IL- 5 responsiveness, but up-regulate IL-3 receptor alpha expression. J Immunol. 2003 Jun 1; 170 (11): 5359-66. (allergic inflammation)

IL-4 / IL-2

• IL-4 synergistically enhances both IL-2and IL-12-induced IFN- {gamma} expression in murine NK cells. Blood. 2003 Mar 13 [Epub ahead of print]

IL-4 / IL-5

• Enhancement of histamine secretion, etc., in mast cells in response to IgE • A Th2-like cytokine response is involved in bullous pemphigoid. the role of IL-4 and IL-5 in the pathogenesis of the disease. Int J Immunopathol Pharmacol. 1999 MayAug; 12 (2): 55-61.

IL-4 / IL-6

IL-4 / IL-10

IL-4 / IL-11

• Synergistic interactions between interleukin-11 and interleukin-4 in support of proliferation of primitive hematopoietic progenitors of mice. Blood. 1991 Sep 15; 78 (6): 1448-51.

IL-4 / IL-12

• Synergistic effects of IL-4 and IL-18 on IL-12-dependent IFN-gamma production by dendritic cells. J Immunol. 2000 Jan 1; 164 (1): 64-71. (increase Th1 / Th2 differentiation) • IL-4 synergistically enhances both IL-2and IL-12-induced IFN- {gamma} expression in murine NK cells. Blood. 2003 Mar 13 [Epub ahead of print]

IL-4 / IL-13

• Summary Interleukin-4 and interleukin-13 signaling connections maps. Science 2003 Jun 6; 300 (5625): 1527-8. (allergy, asthma) • Inhibition of the IL-4 / IL-13 receptor system prevents allergic sensitization without affecting established allergy in a mouse model for allergic asthma. J Allergy Clin Immunol. 2003 Jun; 111 (6): 1361-1369.

IL-4 / IL-16

• (asthma) Interleukin (IL) -4 / IL-9 and exogenous IL-16 induces IL-16 production by BEAS-2B cells, a bronchial epithelial cell line. Immunol cell. 2001 Feb 1; 207 (2): 7580

IL-4 / IL-17

• Interleukin (IL) -4 and IL-17 synergistically stimulate IL-6 secretion in human colonic myofibroblasts. Int J Mol Med. 2002 Nov; 10 (5): 631-4. (Gut inflammation)

IL-4 / IL-24

• IL-24 is expressed by rat and human macrophages. Immunobiology 2002 Jul; 205 (3): 321-34.

IL-4 / IL-25

• Summary New IL-17 family members promote Th1 or Th2 responses in the lung: in vivo function of the novel cytokine IL-25. J Immunol. 2002 Jul 1; 169 (1): 443-53. (allergic inflammation) • Summary Mast cells produces interleukin-25 upon Fc-epsilon RI-mediated activation. Blood. 2003 May 1; 101 (9): 3594-6. Epub 2003 Jan 02. (allergic inflammation)

IL-4 / IFN-γ

• Summary Interleukin 4 induces interleukin 6 production by endothelial cells: synergy with interferon-gamma. Eur J Immunol. 1991 Jan; 21 (1): 97-101.

IL-4 / SCF

• Regulation of human intestinal mast cells by stem cell factor and IL-4. Immunol Rev. 2001 Feb; 179: 57-60. Revision.

IL-5 / IL-3

• Differential regulation of human eosinophil IL-3, IL-5, and GM-CSF receptor alphachain expression by cytokines: IL-3, IL-5, and GM-CSF down-regulate IL-5 receptor alpha expression with loss of IL- 5 responsiveness, but up-regulate IL-3 receptor alpha expression. J Immunol. 2003 Jun 1; 170 (11): 5359-66. (Allergic inflammation,

see summary)

IL-5 / IL-6

IL-5 / IL-13

• Inhibition of allergic airways inflammation and airway hyperresponsiveness in mice by dexamethasone: role of eosinophils, IL-5, eotaxin, and IL-13. J Allergy Clin Immunol. 2003 May; 111 (5): 1049-61.

IL-5 / IL-17

• Interleukin-17 orchestrates the granulocyte influx into airways after allergen inhalation in a mouse model of allergic asthma. Am J Respir Cell Mol Biol. 2003 Jan; 28 (1): 42-50.

IL-5 / IL-25

• Summary New IL-17 family members promote Th1 or Th2 responses in the lung: in vivo function of the novel cytokine IL-25. J Immunol. 2002 Jul 1; 169 (1): 443-53. (allergic inflammation) • Summary Mast cells produces interleukin-25 upon Fcepsilon RI-mediated activation. Blood. 2003 May 1; 101 (9): 3594-6. Epub 2003 Jan 02. (allergic inflammation)

IL-5 / IFN-γ

IL-5 / GM-CSF

• Differential regulation of human eosinophil IL-3, IL-5, and GM-CSF receptor alphachain expression by cytokines: IL-3, IL-5, and GM-CSF down-regulate IL-5 receptor alpha expression with loss of IL- 5 responsiveness, but up-regulate IL-3 receptor alpha expression. J Immunol. 2003 Jun 1; 170 (11): 5359-66. (Allergic inflammation)

IL-6 / IL-10

IL-6 / IL-11

IL-6 / IL-16

• Interleukin-16 stimulates the expression and production of proinflammatory cytokines by human monocytes. Immunology 2000 May; 100 (1): 63-9.

IL-6 / IL-17

• Stimulation of airway mucin gene expression by interleukin (IL) -17 through IL-6 paracrine / autocrine loop. J Biol Chem. 2003 May 9; 278 (19): 17036-43. Epub 2003 Mar 06. (airway inflammation, asthma)

IL-6 / IL-19

• Summary IL-19 induces production of IL-6 and TNF-alpha and results in cell apoptosis through TNF-alpha. J Immunol. 2002 Oct 15; 169 (8): 4288-97.

IL-6 / IFN-g

IL-7 / IL-2

• Interleukin 7 worsens graft-versus-host disease. Blood. 2002 Oct 1; 100 (7): 2642-9.

IL-7 / IL-12

• Synergistic effects of IL-7 and IL-12 on human T cell activation. J Immunol. 1995 May 15; 154 (10): 5093-102.

IL-7 / IL-15

• Interleukin-7 and interleukin-15 regulate the expression of the bcl-2 and c-myb genes in cutaneous T-cell lymphoma cells. Blood. 2001 Nov 1; 98 (9): 2778-83. (growth factor)

IL-8 / IL-11

• Abnormal production of interleukin (IL) -11 and IL-8 in polycythaemia vera. Cytokine 2002 Nov 21; 20 (4): 178-83.

IL-8 / IL-17

• The Role of IL-17 in Joint Destruction. Drug News Perspect. 2002 Jan; 15 (1): 17-23. (arthritis) • Summary Interleukin-17 stimulates the expression of interleukin-8, growth-related oncogene alpha, and granulocyte-colony-stimulating factor by human airway epithelial cells. Am J Respir Cell Mol Biol. 2002 Jun; 26 (6): 748-53. (inflammation of the airways)

IL-8 / GSF

• Interleukin-8: an autocrine / paracrine growth factor for human hematopoietic progenitors acting in synergy with colony stimulating factor-1 to promote monocytemacrophage growth and differentiation. Hematol Exp. 1999 Jan; 27 (1): 28-36.

IL-8 / VGEF

• Intracavitary VEGF, bFGF, IL-8, IL-12 levels in primary and recurrent malignant

glioma J Neurooncol. 2003 May; 62 (3): 297-303.

IL-9 / IL-4

• Anti-interleukin-9 antibody treatment inhibits airway inflammation and hyperreactivity in mouse asthma model. Am J Respir Crit Care Med. 2002 Aug 1; 166 (3): 409-16.

IL-9 / IL-5

• Pulmonary overexpression of IL-9 induces Th2 cytokine expression, leading to immune pathology. J Clin Invest. 2002 Jan; 109 (1): 29-39. • Th2 cytokines and asthma. Interleukin-9 as a therapeutic target for asthma. Respir Res. 2001; 2 (2): 80-4. Epub 2001 Feb 15. Review. • Summary Interleukin-9 enhances interleukin-5 receptor expression, differentiation, and survival of human eosinophils. Blood. 2000 Sep 15; 96 (6): 2163-71 (asthma)

IL-9 / IL-13

• Anti-interleukin-9 antibody treatment inhibits airway inflammation and hyperreactivity in mouse asthma model. Am J Respir Crit Care Med. 2002 Aug 1; 166 (3): 409-16.

• Direct effects of interleukin-13 on epithelial cells cause airway hyperreactivity and mucus overproduction in asthma. Nat Med. 2002 Aug; 8 (8): 885-9.

IL-9 / IL-16

• See IL-4 / IL-16

IL-10 / IL-2

• The interplay of interleukin-10 (IL-10) and interleukin-2 (IL-2) in humoral immune responses: IL-10 synergizes with IL-2 to enhance responses of human B lymphocytes in a mechanism which is different from upregulation of CD25 expression Immunol cell. 1994 Sep; 157 (2): 478-88.

IL-10 / IL-12

IL-10 / TGF-β

• IL-10 and TGF-beta cooperate in the regulatory T cell response to mucosal allergens in normal immunity and specific immunotherapy. Eur J Immunol. 2003 May; 33 (5): 1205-14.

IL-10 / IFN-γ

IL-11 / IL-6

• Interleukin-6 and interleukin-11 support human osteoclast formation by a RANKLindependent mechanism. Bone 2003 Jan; 32 (1): 1-7. (bone resorption in inflammation)

IL-11 / IL-17

• Polarized in vivo expression of IL-11 and IL-17 between acute and chronic skin lesions. J Allergy Clin Immunol. 2003 Apr; 111 (4): 875-81. (allergic dermatitis) • IL-17 promotes bone erosion in murine collagen-induced arthritis through loss of the receptor activator of NF-kappa B ligand / osteoprotegerin balance. J Immunol. 2003 Mar 1; 170 (5): 2655-62.

IL-11 / TGF-β

• Polarized in vivo expression of IL-11 and IL-17 between acute and chronic skin lesions. J Allergy Clin Immunol. 2003 Apr; 111 (4): 875-81. (Allergic dermatitis)

IL-12 / IL-13

• Relationship of Interleukin-12 and Interleukin-13 imbalance with class-specific rheumatoid factors and anticardiolipin antibodies in systemic lupus erythematosus. Clin Rheumatol. 2003 May; 22 (2): 107-11.

IL-12 / IL-17

• Upregulation of interleukin-12 and -17 in active inflammatory bowel disease. Scand J Gastroenterol. 2003 Feb; 38 (2): 180-5.

IL-12 / IL-18

• Synergistic proliferation and activation of natural killer cells by interleukin 12 and interleukin 18. Cytokine. 1999 Nov; 11 (11): 822-30. • Inflammatory Liver Steatosis Caused by IL-12 and IL-18. J Interferon Cytokine Res. 2003 Mar; 23 (3): 155-62.

IL-12 / IL-23

• Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflammation of the brain. Nature 2003 Feb 13; 421 (6924): 744-8. • Summary A unique role for IL-23 in promoting cellular immunity. J Leukoc Biol. 2003 Jan; 73 (1): 49-56. Revision.

IL-12 / IL-27

• Summary IL-27, a heterodimeric cytokine composed of EBI3 and p28 protein, induces proliferation of naive CD4 (+) T cells. Immunity 2002 Jun; 16 (6): 779-90.

IL-12 / IFN-γ

• IL-12 induces IFN-γ expression by B and T lymphocytes as part of immune stimulation.

IL-13 / IL-5

• See IL-5 / IL-13

IL-13 / IL-25

• Summary New IL-17 family members promote Th1 or Th2 responses in the lung: in vivo function of the novel cytokine IL-25. J Immunol. 2002 Jul 1; 169 (1): 443-53. (allergic inflammation) • Summary Mast cells produces interleukin-25 upon Fcepsilon RI-mediated activation. Blood. 2003 May 1; 101 (9): 3594-6. Epub 2003 Jan 02. (allergic inflammation)

IL-15 / IL-13

• Differential expression of interleukins (IL) -13 and IL-15 in ectopic and eutopic endometrium of women with endometriosis and normal fertile women. Am J Reprod Immunol. 2003 Feb; 49 (2): 75-83.

IL-15 / IL-16

• IL-15 and IL-16 overexpression in cutaneous T-cell lymphomas: stage-dependent increase in mycosis fungoides progression. Dermatol Exp. 2000 Aug; 9 (4): 248-51.

IL-15 / IL-17

• Summary IL-17, produced by lymphocytes and neutrophils, is necessary for lipopolysaccharide-induced airway neutrophilia: IL-15 as a possible trigger. J Immunol. 2003 Feb 15; 170 (4): 2106-12. (inflammation of the airways)

IL-15 / IL-21

• IL-21 in Synergy with IL-15 or IL-18 Enhances IFN-gamma Production in Human NK and T Cells. J Immunol. 2003 Jun 1; 170 (11): 5464-9.

IL-17 / IL-23

• Interleukin-23 promotes a distinct CD4 T cell activation state characterized by the production of interleukin-17. J Biol Chem. 2003 Jan 17; 278 (3): 1910-4. Epub 2002 Nov 03

IL-17 / TGF-β

• Polarized in vivo expression of IL-11 and IL-17 between acute and chronic skin lesions. J Allergy Clin Immunol. 2003 Apr; 111 (4): 875-81. (Allergic dermatitis)

IL-18 / IL-12

• Synergistic proliferation and activation of natural killer cells by interleukin 12 and interleukin 18. Cytokine. 1999 Nov; 11 (11): 822-30. • Summary Inhibition of in vitro immunoglobulin production by IL-12 in murine chronic graftvs.host disease: synergism with IL-18. Eur J Immunol. 1998 Jun; 28 (6): 2017-24.

IL-18 / IL-21

• IL-21 in Synergy with IL-15 or IL-18 Enhances IFN-gamma Production in Human NK and T Cells. J Immunol. 2003 Jun 1; 170 (11): 5464-9.

IL-18 / TGF-β

• Interleukin 18 and transforming growth factor beta1 in the serum of patients with Graves ’ophthalmopathy treated with corticosteroids. Int Immunopharmacol. 2003 Apr; 3 (4): 549-52.

IL-18 / IFN-γ

Anti-TNF ALPHA / anti-CD4

• Synergistic therapeutic effect in DBA / 1 arthritic mice.

Annex 3: Oncology combinations

Diana

Disease Pair with

CD89 *

Use as a cytotoxic cell recruiter everybody

CD19

B lymphocyte lymphomas HLA-DR CD5

HLA-DR

B lymphocyte lymphomas CD89 CD19 CD5

CD38

Multiple myeloma CD138 CD56 HLA-DR

CD138

Multiple myeloma CD38 CD56 HLA-DR

CD138

Lung cancer CD56 CEA

CD33

Acute myeloid lymphoma CD34 HLA-DR

CD56

Lung cancer CD138 CEA

CEA

Various carcinomas MET receiver

VEGF

Various carcinomas MET receiver

VEGF receiver

Various carcinomas MET receiver

IL-13

Asthma / lung inflammation IL-4 IL-5 Eotaxin (s) MDC TARC TNFα IL-9 EGFR CD40L IL-25 MCP-1 TGFβ

IL-4

Asthma IL-13

IL-5 Eotaxin (s) MDC TARC TNFα IL-9 EGFR CD40L IL-25 MCP-1 TGFβ

Eotaxin

Asthma IL-5 Eotaxin-2 Eotaxin-3

EGFR

Cancer HER2 / neu HER3 HER4

HER2

Cancer HER3 HER4

TNFR1

RA / Crohn's disease IL-1R IL-6R IL-18R

TNFα

RA / Crohn's disease IL-1α / β IL-6 IL-18 ICAM-1 IL-15 IL-17

IL-1R

RA / Crohn's disease IL-6R IL-18R

IL-18R

RA / Crohn's disease IL-6R

Annex 4

Data summary

DIANA

dAb Equilibrium dissociation constant (Kd = Koff / Kon) Koff IC50 for ligand assay DN50 for cell-based neutrality assay

TAR1

TAR1 monomers 300 nM to 5 pM (e.g., 3 x 10-7 to 5 x 10 -12), preferably 50 nM to 20 pM 5 x 10-1 to 1 x 10-7 500 nM to 100 pM 500 nM at 50 pM

TAR1 dimers

As TAR1 monomer As TAR1 monomer As TAR1 monomer As TAR1 monomer

TAR1 trimers

As TAR1 monomer As TAR1 monomer As TAR1 monomer As TAR1 monomer

TAR1-5

TAR1-27

TAR1-5-19 monomer

30 nM

TAR1-5-19 homodimer

With linker group (Gly4 Ser) 3 = 20 nM With linker group (Gly4 Ser) 5 = 2 nM With linker group (Gly4 Ser) 7 = 10 nM In Fab format = 1 nM

= 30 nM = 3 nM = 15 nM

TAR1-5-19 heterodimers

With linker group (Gly4 Ser) n TAR1-5-19 d2 = 2 nMTAR1-5-19 d3 = 8 nMTAR1-5-19 d4 = 2-5 nMTAR1-5-19 d5 = 8 nM in FabTAR1-5-19CH format d1CK = 6 nMTAR1-5-19CK d1CH = 6 nMTAR1-5-19CH d2CK = 8 nM = 12 nM = 10 nM

TAR1-5-19CH d3CK = 3 nM

= 12 nM

TAR1-5 heterodimers

With linker group (Gly4 Ser) n TAR1-5d1 = 30 nMTAR1-5d2 = 50 nMTAR1-5d3 = 300 nMTAR1-5d4 = 3 nMTAR1-5d5 = 200 nMTAR1-5d6 = 100 nM in FabTAR1-5CH d2CK format = 30 nMTAR1-5CK d3CH = 100 nM = 60 nM

TAR1-5-19 homodimer

0.3 nM 3-10 nM (e.g., 3 nM)

TAR2

TAR2 monomers As TAR1 monomer As TAR1 monomer 500 nM to 100 pM 500 nM at 50 pM

TAR2-10

TAR2-5

Albuminic

Monómeros anti-SA 1 nM to 500 M, preferably 100 nM to 10 M in a specific double format, target affinity is 1 to 100,000 x affinity dAb SA, p. ex. 100 pM (target) and 10 M affinity SA. 1 nM to 500 M, preferably 100 nM to 10 M in a specific double formation, the target affinity is 1 to 100,000 x affinity dAb SA, p. ex. 100 pM (target) and 10 M affinity SA.

MSA-16

200 nM

MSA-26

70 nM

Appendix 1 - Sequence listing

HEL4

TAR2-5

TAR1-5-19 TAR2-10

DNA and amino acid sequences of dAb of TAR-1. DAbs 1, 2 and 3 are the associated sequences of dimers 1-6 of TAR1-5. DAb1 is the dAb associated with dimers 1, 5 and 6 The second dAb are the same, but the linker lengths are different, also, dAb2 is the dAb

5 associated with dimers 2 and 3, dAb3 is the dAb associated with dimer 4. * indicates the presence of an amber stop codon. The sequence homology between TAR1-5, TAR1-5-19 and the only second dAbs is 88 %

TAR1-5-19

TAR1-5

Monomer -dAb1: Associated DAb in TAR1-5d1 (3U linker), d5 (5U linker), d6 (7U linker)

Monomer -dAb2: associated DAb in TAR1-5d2 (3U linker), d3 (5U linker)

Monomer -dAb3: associated DAb in TAR1-5d4 (5U linker)

TAR1-27 and associated amino acid and dAb DNA sequences (* indicates the presence of an amber stop codon). The sequence homology between TAR1-27 and the second single dAbs is 90.4%.

TAR1-27

Associated dAb monomer in TAR1-27d1 (3U linker) Associated dAb monomer in TAR1-27d2 (3U linker)

Associated dAb monomer in TAR1-27d7 (3U linker) Associated dAb monomer in TAR1-27d8 (3U linker)

Associated dAb monomer in TAR1-27d12 (3U linker) Associated dAb monomer in TAR1-27d16 (3U linker)

Associated dAb monomer in TAR1-27d23 (5U linker) Associated dAb monomer in TAR1-27d30 (7U linker)

Associated dAb monomer in TAR1-27d31 (7U linker) Associated dAb monomer in TAR1-27d36 (7U linker)

Associated dAb monomer in TAR1-27d37 (7U linker) Associated dAb monomer in TAR1-27d39 (7U linker)

Amino acid and DNA sequences of TAR2 dAb (* indicates the presence of an amber stop codon) The sequence homology between TAR2h-5 and the second single dAbs is 83.5%

TAR2h-5

Associated dAb monomer in TAR2h-5d1 (3U linker) Associated dAb monomer in TAR2h-5d3 (3U linker)

Associated dAb monomer in TAR2h-5d4 (3U linker) Associated dAb monomer in TAR2h-5d5 (3U linker)

Associated dAb monomer in TAR2h-5d6 (3U linker) Associated dAb monomer in TAR2h-5d7 (3U linker)

Associated dAb monomer in TAR2h-5d8 (3U linker) Associated dAb monomer in TAR2h-5d9 (3U linker)

Associated dAb monomer in TAR2h-5d10 (5U linker) Associated dAb monomer in TAR2h-5d11 (5U linker)

Associated dAb monomer in TAR2h-5d12 (5U linker) Associated dAb monomer in TAR2h-5d13 (5U linker)

Claims (5)

1. A ligand comprising an individual variable domain of antibodies, wherein said individual variable domain of antibodies comprises the amino acid sequence of dAb7h11 having the amino acid sequence depicted in Figure 24, Sheet 1, or an amino acid sequence that is identical to it at 5 less in 95%, where the beta T half-life of the individual variable domain in cynomolgus is at least 50% of the beta T half-life of cynomolgus serum albumin in cynomolgus.

2.

A ligand according to claim 1, wherein the individual variable domain of antibodies binds to the serum albumin of rat, mouse, human and Cynomolgus.

3.

A ligand according to any preceding claim, wherein said individual variable domain of antibodies

10 specifically binds to human serum albumin and non-human serum albumin with Kd values within the limits of a factor of 10 with each other. 4. A ligand according to any preceding claim further comprising a biologically active polypeptide. 5. A ligand according to any preceding claim wherein said individual variable domain of antibodies 15 is conjugated to a drug. FIG. 13 FIG. 14 FIG. fifteen FIG. 16 FIG. 19 FIG. twenty Figure 24. Antibody sequences of AlbudAb clones identified by phage selection Figure 24 (continued) Figure 25. Alignments of the three domains of human serum albumin. You can clearly see the conservation of cysteine residues.